Search

The earlier prediction of heart diseases and appropriate treatment are important for preventing cardiac failure complications and reducing the mortality rate. The traditional prediction and classification approaches have resulted in a minimum rate of prediction accuracy and hence to overcome the pitfalls in existing systems, the present research is aimed to perform the prediction of heart diseases with quantum learning. When quantum learning is employed in ML (Machine Learning) and DL (Deep Learning) algorithms, complex data can be performed efficiently with less time and a higher accuracy rate. Moreover, the proposed ML and DL algorithms possess the ability to adapt to predictions with alterations in the dataset integrated with quantum computing that provides robustness in the earlier detection of chronic diseases. The Cleveland heart disease dataset is being pre-processed for the checking of missing values to avoid incorrect predictions and also for improvising the rate of accuracy. Further, SVM (Support Vector Machine), DT (Decision Tree) and RF (Random Forest) are used to perform classification. Finally, disease prediction is performed with the proposed instance-based quantum ML and DL method in which the number of qubits is computed with respect to features and optimized with instance-based learning. Additionally, a comparative assessment is provided for quantifying the differences between the standard classification algorithms with quantum-based learning in order to determine the significance of quantum-based detection in heart failure. From the results, the accuracy of the proposed system using instance-based quantum DL and instance-based quantum ML is found to be 98% and 83.6% respectively. [ABSTRACT FROM AUTHOR]

Diagnosis of adamantinomatous craniopharyngioma (ACP) is predominantly determined through invasive pathological examination of a neurosurgical biopsy specimen. Clinical experts can distinguish ACP from Magnetic Resonance Imaging (MRI) with an accuracy of 86%, and 9% of ACP cases are diagnosed this way. Classification using deep learning (DL) provides a solution to support a non-invasive diagnosis of ACP through neuroimaging, but it is still limited in implementation, a major reason being the lack of predictive uncertainty representation. We trained and tested a DL classifier on preoperative MRI from 86 suprasellar tumor patients across multiple institutions. We then applied a Bayesian DL approach to calibrate our previously published ACP classifier, extending beyond point-estimate predictions to predictive distributions. Our original classifier outperforms random forest and XGBoost models in classifying ACP. The calibrated classifier underperformed our previously published results, indicating that the original model was overfit. Mean values of the predictive distributions were not informative regarding model uncertainty. However, the variance of predictive distributions was indicative of predictive uncertainty. We developed an algorithm to incorporate predicted values and the associated uncertainty to create a classification abstention mechanism. Our model accuracy improved from 80.8% to 95.5%, with a 34.2% abstention rate. We demonstrated that calibration of DL models can be used to estimate predictive uncertainty, which may enable clinical translation of artificial intelligence to support non-invasive diagnosis of brain tumors in the future. [ABSTRACT FROM AUTHOR]

Malware attacks on the Android platform are rapidly increasing due to the high consumer adoption of Android smartphones. Advanced technologies have motivated cyber-criminals to actively create and disseminate a wide range of malware on Android smartphones. The researchers have conducted numerous studies on the detection of Android malware, but the majority of the works are based on the detection of generic Android malware. The detection based on malware categories will provide more insights about the malicious patterns of the malware. Therefore, this paper presents a detection solution for different Android malware categories, including adware, banking, SMS malware, and riskware. In this paper, a novel Huffman encoding-based feature vector generation technique is proposed. The experiments have proved that this novel approach significantly improves the efficiency of the detection model. This method makes use of system call frequencies as features to extract malware's dynamic behavior patterns. The proposed model was evaluated using machine learning and deep learning methods. The results show that the proposed model with the Random Forest classifier outperforms some existing methodologies with a detection accuracy of 98.70%. [ABSTRACT FROM AUTHOR]

Kicks can lead to well control risks during petroleum drilling, and even more serious kicks may lead to serious casualties, which is the biggest threat factor affecting the safety in the process of petroleum drilling. Therefore, how to detect kicks early and efficiently has become a focus practical problem. Traditional machine learning models require a large amount of labeled data, such kicked sample, and it is difficult to label data, which requires a lot of labor and time. To address the above issues, the deep forest is extended to a transfer learning model to improve the generalization ability. In this paper, a transfer learning model is built to detect kicks early. The source domain model adopts the deep forest model. Deep forest is an ensemble learning model with a hierarchical structure similar to deep learning. Each layer contains a variety of random forests. It is an integration of the model in depth and breadth. In the case of a small sample size (20–60 min), kick can be identified 10 min in advance. The deep forest model is established as the source domaining model, and a cascade forest is added at the last layer according to the transfer learning algorithm to form the classification model of this paper. The experimental results show that the kick prediction accuracy of the model is 80.13% by a confusion matrix. In the target domain, the proposed model performs better than other ensemble learning algorithms, and the accuracy is 5% lower than other SOTA transfer learning algorithms. [ABSTRACT FROM AUTHOR]

Gridded gross domestic product (GDP) data are a crucial land surface parameter for many geoscience applications. Recently, machine learning approaches have become powerful tools in generating gridded GDP data. However, most machine learning approaches for gridded GDP estimation seldom consider the geographical properties of input variables. Therefore, in this study, a geographically weighted stacking ensemble learning approach was developed to generate gridded GDP data. Three algorithms—random forest, XGBoost, and LightGBM—were used as base models, and the linear regression in stacking ensemble learning was replaced by geographically weighted regression to locally fuse the three predictions. A case study was conducted in China to demonstrate the effectiveness of the proposed approach. The results showed that the proposed GDP downscaling approach outperformed the three base models and traditional stacking ensemble learning. Meanwhile, it had good predictive power on county-level GDP test data with R2 of 0.894, 0.976, and 0.976 for the primary, secondary, and tertiary sectors, respectively. Moreover, the predicted 1 km gridded GDP data had a high accuracy (R2 = 0.787) when evaluated by town-level GDP data. Hence, the proposed GDP downscaling approach provides a valuable option for generating gridded GDP data. The generated 1 km gridded GDP data of China from 2020 are of great significance for other applications. [ABSTRACT FROM AUTHOR]

Against the background of the ongoing atmospheric warming, the glacial lakes that are nourished and expanded in High Mountain Asia pose growing risks of glacial lake outburst floods (GLOFs) hazards and increasing threats to the downstream areas. Effectively extracting the area and consistently monitoring the dynamics of these lakes are of great significance in predicting and preventing GLOF events. To automatically extract the lake areas, many deep learning (DL) methods capable of capturing the multi-level features of lakes have been proposed in segmentation and classification tasks. However, the portability of these supervised DL methods need to be improved in order to be directly applied to different data sources, as they require laborious effort to collect the labeled lake masks. In this work, we proposed a simple glacial lake extraction model (SimGL) via weakly-supervised contrastive learning to extend and improve the extraction performances in cases that lack the labeled lake masks. In SimGL, a Siamese network was employed to learn similar objects by maximizing the similarity between the input image and its augmentations. Then, a simple Normalized Difference Water Index (NDWI) map was provided as the location cue instead of the labeled lake masks to constrain the model to capture the representations related to the glacial lakes and the segmentations to coincide with the true lake areas. Finally, the experimental results of the glacial lake extraction on the 1540 Landsat-8 image patches showed that our approach, SimGL, offers a competitive effort with some supervised methods (such as Random Forest) and outperforms other unsupervised image segmentation methods in cases that lack true image labels. [ABSTRACT FROM AUTHOR]

Runoff forecasting is important for water resource management. Although deep learning models have substantially improved the accuracy of runoff prediction, the temporal and feature dependencies between rainfall–runoff time series elements have not been effectively exploited. In this work, we propose a new hybrid deep learning model to predict hourly streamflow: SA-CNN-LSTM (self-attention, convolutional neural network, and long short-term memory network). The advantages of CNN and LSTM in terms of data extraction from time series data are combined with the self-attention mechanism. By considering interdependences of the rainfall–runoff sequence between timesteps and between features, the prediction performance of the model is enhanced. We explored the performance of the model in the Mazhou Basin, China; we compared its performance with the performances of LSTM, CNN, ANN (artificial neural network), RF (random forest), SA-LSTM, and SA-CNN. Our analysis demonstrated that SA-CNN-LSTM demonstrated robust prediction with different flood magnitudes and different lead times; it was particularly effective within lead times of 1–5 h. Additionally, the performance of the self-attention mechanism with LSTM and CNN alone, respectively, was improved at some lead times; however, the overall performance was unstable. In contrast, the hybrid model integrating CNN, LSTM, and the self-attention mechanism exhibited better model performance and robustness. Overall, this study considers the importance of temporal and feature dependencies in hourly runoff prediction, then proposes a hybrid deep learning model to improve the performances of conventional models in runoff prediction. [ABSTRACT FROM AUTHOR]

Insufficient data is always a challenge for developing an accurate machine learning or deep learning model in manufacturing processes, especially in tool wear monitoring under varied cutting conditions. This paper presents a Random Forest model for predicting tool wear under varied cutting conditions as well as studies extracted signal features. The Random Forest algorithm was chosen as the machine learning model, rather than the novel deep learning model. This was due to the feature importance investigation, which was embedded in the Random Forest algorithm, thereby making it easier to study the physical meanings of signal features. The frequency domain signals were rearranged as features related to spindle speeds and machine tool structure based on domain knowledge. This is the first paper to rearrange the frequency domain signals for observing the physical meanings of selected features. When data normalization was adopted, frequency domain signals related to spindle speeds were excluded from important features. Only spectrum energy related to structure vibration and time domain signals were important features. Data normalization enhanced the weighting of structure vibration features in a machine learning model. This study showed that feature normalization made the machine learning model more adaptable to different cutting conditions. Furthermore, prediction accuracy for cutting condition of spindle speed = 42,000 rpm and feed = 1.5 μm/rev (lowest prediction accuracy among cutting tests in this study) showed an increase from 68.0 to 84.1%. In addition, spindle speed had a more significant effect than feed on classification accuracy in tool wear monitoring based on experimental results. As a result, at least two data sets of the same spindle speed as in tool wear prediction were recommended to be used for model training. When there were at least two data sets in training data with the same spindle speed as in testing data, the study showed prediction accuracies were greater than 75% without data normalization and 81% with data normalization. [ABSTRACT FROM AUTHOR]

Fungal species identification from metagenomic data is a highly challenging task. Internal Transcribed Spacer (ITS) region is a potential DNA marker for fungi taxonomy prediction. Computational approaches, especially deep learning algorithms, are highly efficient for better pattern recognition and classification of large datasets compared to in silico techniques such as BLAST and machine learning methods. Here in this study, we present CNN_FunBar, a convolutional neural network-based approach for the classification of fungi ITS sequences from UNITE+INSDC reference datasets. Effects of convolution kernel size, filter numbers, k-mer size, degree of diversity and category-wise frequency of ITS sequences on classification performances of CNN models have been assessed at all taxonomic levels (species, genus, family, order, class and phylum). It is observed that CNN models can produce >93% average accuracy for classifying ITS sequences from balanced datasets with 500 sequences per category and 6-mer frequency features at all levels. The comparative study has revealed that CNN_FunBar can outperform machine learning-based algorithms (SVM, KNN, Naïve-Bayes and Random Forest) as well as existing fungal taxonomy prediction software (funbarRF, Mothur, RDP Classifier and SINTAX). The present study will be helpful for fungal taxonomy classification using large metagenomic datasets. [ABSTRACT FROM AUTHOR]

Recently, it has proven difficult to make an immediate remote diagnosis of any coronary illness, including heart disease, diabetes, etc. The drawbacks of cloud computing infrastructures, such as excessive latency, bandwidth, energy consumption, security, and privacy concerns, have lately been addressed by Fog computing with IoT applications. In this study, an IoT-Fog-Cloud integrated system, called a Fog-empowered framework for real-time analysis in heart patients using ENsemble Deep learning (FRIEND), has been introduced that can instantaneously facilitate remote diagnosis of heart patients. The proposed system was trained on the combined dataset of Long-Beach, Cleveland, Switzerland, and Hungarian heart disease datasets. We first tested the model with eight basic ML approaches, including the decision tree, logistic regression, random forest, naive Bayes, k-nearest neighbors, support vector machine, AdaBoost, and XGBoost approaches, and then applied ensemble methods including bagging classifiers, weighted averaging, and soft and hard voting to achieve enhanced outcomes and a deep neural network, a deep learning approach, with the ensemble methods. These models were validated using 16 performance and 9 network parameters to justify this work. The accuracy, PPV, TPR, TNR, and F1 scores of the experiments reached 94.27%, 97.59%, 96.09%, 75.44%, and 96.83%, respectively, which were comparatively higher when the deep neural network was assembled with bagging and hard-voting classifiers. The user-friendliness and the inclusion of Fog computing principles, instantaneous remote cardiac patient diagnosis, low latency, and low energy consumption, etc., are advantages confirmed according to the achieved experimental results. [ABSTRACT FROM AUTHOR]

The Machine Learning/Deep Learning (ML/DL) forecasting model has helped stakeholders overcome uncertainties associated with renewable energy resources and time planning for probable near-term power fluctuations. Nevertheless, the effectiveness of long-term forecasting of renewable energy resources using an existing ML/DL model is still debatable and needs additional research. Considering the constraints inherent in current empirical or physical-based forecasting models, the study utilizes ML/DL models to provide long-term predictions for solar power production. This study aims to examine the efficacy of several existing forecasting models. The study suggests approaches to enhance the accuracy of long-term forecasting of solar power generation for a case study power plant. It summarizes and compares the statistical model (ARIMA), ML model (SVR), DL models (LSTM, GRU, etc.), and ensemble models (RF, hybrid) with respect to long-term prediction. The performances of the univariate and multivariate models are summarized and compared based on their ability to accurately predict solar power generation for the next 1, 3, 5, and 15 days for a 100-kW solar power plant in Lubbock, TX, USA. Conclusions are drawn predicting the accuracy of various model changes with variation in the prediction time frame and input variables. In summary, the Random Forest model predicted long-term solar power generation with 50% better accuracy over the univariate statistical model and 10% better accuracy over multivariate ML/DL models. [ABSTRACT FROM AUTHOR]

High-resolution mapping of PM2.5 concentration over Tehran city is challenging because of the complicated behavior of numerous sources of pollution and the insufficient number of ground air quality monitoring stations. Alternatively, high-resolution satellite Aerosol Optical Depth (AOD) data can be employed for high-resolution mapping of PM2.5. For this purpose, different data-driven methods have been used in the literature. Recently, deep learning methods have demonstrated their ability to estimate PM2.5 from AOD data. However, these methods have several weaknesses in solving the problem of estimating PM2.5 from satellite AOD data. In this paper, the potential of the deep ensemble forest method for estimating the PM2.5 concentration from AOD data was evaluated. The results showed that the deep ensemble forest method with R 2 = 0.74 gives a higher accuracy of PM2.5 estimation than deep learning methods ( R 2 = 0.67 ) as well as classic data-driven methods such as random forest ( R 2 = 0.68 ). Additionally, the estimated values of PM2.5 using the deep ensemble forest algorithm were used along with ground data to generate a high-resolution map of PM2.5. Evaluation of produced PM2.5 map revealed the good performance of the deep ensemble forest for modeling the variation of PM2.5 in the city of Tehran. [ABSTRACT FROM AUTHOR]

The evergrowing diversity of encrypted and anonymous network traffic makes network management more formidable to manage the network traffic. An intelligent system is essential to analyse and identify network traffic accurately. Network management needs such techniques to improve the Quality of Service and ensure the flow of secure network traffic. However, due to the usage of non‐standard ports and encryption of data payloads, the classical port‐based and payload‐based classification techniques fail to classify the secured network traffic. To solve the above‐mentioned problems, this paper proposed an effective deep learning‐based framework employed with flow‐time‐based features to predict heterogeneous secure network traffic best. The state‐of‐the‐art machine learning strategies (C4.5, random forest, and K‐nearest neighbour) are investigated for comparison. The proposed 1D‐CNN model achieved higher accuracy in classifying the heterogeneous secure network traffic. In the next step, the proposed deep learning model characterises the major categories (virtual private network traffic, the onion router network traffic, and plain encrypted network traffic) into several application types. The experimental results show the effectiveness and feasibility of the proposed deep learning framework, which yields improved predictive power compared to the state‐of‐the‐art machine learning techniques employed for secure network traffic analysis. [ABSTRACT FROM AUTHOR]

Industrial refrigeration systems typically exhibit a higher energy consumption rate compared to other systems. It is a very important system, and for the brewing industry, it is at the heart of brewing as, in each process, the temperature of the beer must be controlled. Therefore, a health monitoring or faults detection system is a necessary tool for predictive and preventive maintenance to avoid faults in the system. This work presents machine learning models for faults prediction of the refrigeration system in a brewing firm. The raw data of processing parameters are collected, while the system health for each data set is identified and used for machine learning model training. Three operation cases based on the actual operation of the brewery's refrigeration system are analyzed in this work. Several machine learning models including Naïve Bayes (NB), Generalized Linear Models (GLMs), Logistic Regression (LR), Fast Large Margin (FLM), Deep Learning (DL), Decision Tree (DT), Random Forest (RF), Gradient Boosted Trees (GBT), and Support Vector Machine (SVM) are applied for refrigeration system fault detection in the three operation cases while their performances are investigated. The numerical simulation is performed based on RapidMiner Studio. According to the experiment of the three operation cases, the deep learning model was found to be the most accurate and required the least amount of time for the analysis. The accuracy percentages are 86.7%, 90.2%, and 82.5%, while the running times are 15 seconds, 27 seconds, and 15 seconds, respectively. This work can be considered as the baseline for future studies on applied machine learning models for fault detection in refrigeration systems. [ABSTRACT FROM AUTHOR]

As the most widely used storage device today, hard disks are efficient and convenient, but the damage incurred in the event of a failure can be very significant. Therefore, early warnings before hard disk failure, allowing the stored content to be backed up and transferred in advance, can reduce many losses. In recent years, an endless stream of research on the prediction of hard disk failure prediction has emerged. The detection accuracy of various methods, from basic machine learning models, such as decision trees and random forests, to deep learning methods, such as BP neural networks and recurrent neural networks, has also been improving. In this paper, based on the idea of blending ensemble learning, a novel failure prediction method combining machine learning algorithms and neural networks is proposed on the publicly available BackBlaze hard disk datasets. The failure prediction experiment is conducted only with S.M.A.R.T., that is, the learned characteristics collected by self-monitoring analysis and reporting technology, which are internally counted during the operation of the hard disk. The experimental results show that this ensemble learning model is able to outperform other independent models in terms of evaluation criterion based on the Matthews correlation coefficient. Additionally, through the experimental results on multiple types of hard disks, an ensemble learning model with high performance on most types of hard disks is found, which solves the problem of the low robustness and generalization of traditional machine learning methods and proves the effectiveness and high universality of this method. [ABSTRACT FROM AUTHOR]

As agriculture has an increasing impact on the environment, new techniques can help meet future food needs while maintaining or reducing the environmental footprint. Those techniques must incorporate a range of sensing, communication, and data analysis technologies to make informed management decisions, such as those related to the use of water, fertilizer, pesticides, seeds, fuel, labor, etc., to help increase crop production and reduce water and nutrient losses, as well as negative environmental impacts. In this study, a Bidirectional Long Short-Term Memory (BiLSTM) model was trained on real data from Internet of Things sensors in a vineyard located in the Douro wine-growing region, from 2018–2021, to evaluate the ability of this model to predict the Fraction of Transpirable Soil Water (FTSW). The model uses historical data, including reference evapotranspiration, relative humidity, vapor pressure deficit, and rainfall, and outputs the FTSW for periods of one, three, five, and seven days. The model achieved an RMSE between 8.3% and 16.6% and an R 2 -score between 0.75 and 0.93. The model was validated on an independent dataset collected in 2002–2004 from a different vineyard located in the Lisbon wine-growing region, Portugal, and achieved an R 2 -score of 87% and an RMSE of 10.36%. Finally, the performance of the FTSW in the vineyard prediction model was compared with that of the Random Forest model, support vector regression, and linear regression. The results showed that BiLSTM performed better than the RF model on the unseen data, and the BiLSTM model can be considered a suitable model for the accurate prediction of the FTSW. [ABSTRACT FROM AUTHOR]

Tree species surveys are crucial in forest resource management and can provide references for forest protection policymakers. Traditional tree species surveys in the field are labor-intensive and time-consuming. In contrast, airborne LiDAR technology is highly capable of penetrating forest vegetation; it can be used to quickly obtain three-dimensional information regarding vegetation over large areas with a high level of precision, and it is widely used in the field of forestry. At this stage, most studies related to individual tree species classification focus on traditional machine learning, which often requires the combination of external information such as hyperspectral cameras and has difficulty in selecting features manually. In our research, we directly processed the point cloud from a UAV LiDAR system without the need to voxelize or grid the point cloud. Considering that relationships between disorder points can be effectively extracted using Transformer, we explored the potential of a 3D deep learning algorithm based on Transformer in the field of individual tree species classification. We used the UAV LiDAR data obtained in the experimental forest farm of Northeast Forestry University as the research object, and first, the data were preprocessed by being denoised and ground filtered. We used an improved random walk algorithm for individual tree segmentation and made our own data sets. Six different 3D deep learning neural networks and random forest algorithms were trained and tested to classify the point clouds of three tree species. The results show that the overall classification accuracy of PCT based on Transformer reached up to 88.3%, the kappa coefficient reached up to 0.82, and the optimal point density was 4096, which was slightly higher than that of the other deep learning algorithms we analyzed. In contrast, the overall accuracy of the random forest algorithm was only 63.3%. These results show that compared with the commonly used machine learning algorithms and a few algorithms based on multi-layer perceptron, Transformer-based networks provide higher accuracy, which means they can provide a theoretical basis and technical support for future research in the field of forest resource supervision based on UAV remote sensing. [ABSTRACT FROM AUTHOR]

In Mobile Performance Marketing (MPM), monetary compensation only occurs when an advertisement results in a conversion (e.g., sale of a product or service). In this work, we propose an intelligent decision support system (IDSS) to automatically select mobile marketing campaigns for users. The IDSS is based on a computationally efficient mobile user conversion prediction model that assumes a novel Percentage Categorical Pruning (PCP) categorical preprocessing and an online deep multilayer perceptron (MLP) reuse model (MLPr). Using private (nonpublicly available) business MPM data provided by a marketing company, the MLPr model outperformed an offline multilayer perceptron and a logistic regression, obtaining a high quality class discrimination when applied to sampled (85% to 92%) and complete (90% to 94%) data. In addition, the MLPr compared favorably with other machine learning (ML) models (e.g., Random Forest, XGBoost), as well as with other deep neural networks (e.g., diamond shaped). Moreover, we designed two strategies (A — best campaign selection; and B — random selection among the top candidate campaigns) to build the IDSS, in which the predictive deep learning model is used to perform a real-time selection of advertisement campaigns for mobile users. Using recently collected big data (with millions of redirect events) from a worldwide MPM company, we performed a realistic IDSS evaluation that considered three criteria: response time, potential profit and advertiser diversity. Overall, competitive results were achieved by the IDSS B strategy when compared with the current marketing company ad assignment method. [ABSTRACT FROM AUTHOR]

Introduction: Medical tourism, one of the most profitable industries, has been growing rapidly in recent years. Especially Turkey, which has a high ranking among medical travel destinations, has some advantages that can become preferable for international patients. This study is among the first few studies which examine affecting factors in patients' medical travel destination choices with Data Mining techniques. Methods: The data were obtained from patients who came to Ankara from abroad for treatment in May 2015 through a question-naire. Cross-industry Standard Process for data mining, known as the CRISP-DM method, is used in this study. After cleaning out the missing data, the models were created using classification algorithms. Results: Models including Generalized Linear Model, Deep Learning, Decision Tree, Random Forest, Gradient Boosted Trees, and Support Vector Machine (SVM) were compared, and SVM reached the best performance with 0.2% Relative Error, 0.014 Root Mean Squared Error and 0.998 Correlation. As a result of the SVM model, effective attributes in patients' satisfaction level include low price advantage, advertisement, doctors with high-quality education, trained assistant staff, relatives living in Turkey, and high technology of medical equipment, respectively. Conclusion: Special attention should be paid to these factors in developing plans and policies for the health tourism sector. However, the importance of related socio-demographic variables was indicated in detail. Eventually, some suggestions were presented to improve the weaknesses in the health tourism sector. [ABSTRACT FROM AUTHOR]

Breast cancer (BrCa) is the most common disease in women worldwide. Classifying the BrCa image is extremely important for finding BrCa at an earlier stage and monitoring BrCa during treatment. The computer-aided detection methods have been used to interpret BrCa and improve the detection of BrCa during the screening and treatment stages. However, if a new BrCa image is generated for the treatment, it will not classify correctly. The main objective of this research is to classify the BrCa images for newly generated images. The model performs preprocessing, segmentation, feature extraction, and classification. In preprocessing, a hybrid median filtering (HMF) is used to eliminate the noise in the images. The contrast of the images is enhanced using quadrant dynamic histogram equalization (QDHE). Then, ROI segmentation is performed using the USE-Net deep learning model. The CaffeNet model is used for feature extraction on the segmented images, and finally, classification is made using the improved random forest (IRF) with extreme gradient boosting (XGB). The model obtained 97.87% accuracy, 98.45% sensitivity, 95.24% specificity, 98.96% precision, and 98.70% f1-score for ultrasound images. The model gives 98.31% accuracy, 99.29% sensitivity, 90.20% specificity, 98.82% precision, and 99.05% f1-score for mammogram images. [ABSTRACT FROM AUTHOR]

The portable executable header (PEH) information is commonly used as a feature for malware detection systems to train and validate machine learning (ML) or deep learning (DL) classifiers. We propose to extract the deep features from the PEH information through hidden layers of a feed-forward deep neural network (FFDNN). The extraction of deep features of hidden layers represents the dataset with a better generalization for malware detection. While feeding the deep feature of one hidden layer to the succeeding layer, the Gaussian error linear unit (GeLU) activation function is applied. The FFDNN is trained with the GeLU activation function using the deep features of individual layers as well as concatenated deep features of all hidden layers. Similarly, the ML classifiers are also trained and validated in with individual layer deep features and concatenated features. Three highly effective ML classifiers, random forest (RF), support vector machine (SVM), and k-nearest neighbour (k-NN) have been investigated. The performance of the proposed model is demonstrated using a statically significant large dataset. The obtained results are interesting and encouraging in terms of classification accuracy. The classification accuracy reaches 99.15% with the internal discriminative deep feature for the proposed FFDNN-ML classifier with the GeLU activation function. [ABSTRACT FROM AUTHOR]

Artificial intelligence and especially deep learning methods have achieved outstanding results for various applications in the past few years. Pain recognition is one of them, as various models have been proposed to replace the previous gold standard with an automated and objective assessment. While the accuracy of such models could be increased incrementally, the understandability and transparency of these systems have not been the main focus of the research community thus far. Thus, in this work, several outcomes and insights of explainable artificial intelligence applied to the electrodermal activity sensor data of the PainMonit and BioVid Heat Pain Database are presented. For this purpose, the importance of hand-crafted features is evaluated using recursive feature elimination based on impurity scores in Random Forest (RF) models. Additionally, Gradient-weighted class activation mapping is applied to highlight the most impactful features learned by deep learning models. Our studies highlight the following insights: (1) Very simple hand-crafted features can yield comparative performances to deep learning models for pain recognition, especially when properly selected with recursive feature elimination. Thus, the use of complex neural networks should be questioned in pain recognition, especially considering their computational costs; and (2) both traditional feature engineering and deep feature learning approaches rely on simple characteristics of the input time-series data to make their decision in the context of automated pain recognition. [ABSTRACT FROM AUTHOR]

The prediction of drug-target interactions (DTIs) is a vital step in drug discovery. The success of machine learning and deep learning methods in accurately predicting DTIs plays a huge role in drug discovery. However, when dealing with learning algorithms, the datasets used are usually highly dimensional and extremely imbalanced. To solve this issue, the dataset must be resampled accordingly. In this paper, we have compared several data resampling techniques to overcome class imbalance in machine learning methods as well as to study the effectiveness of deep learning methods in overcoming class imbalance in DTI prediction in terms of binary classification using ten (10) cancer-related activity classes from BindingDB. It is found that the use of Random Undersampling (RUS) in predicting DTIs severely affects the performance of a model, especially when the dataset is highly imbalanced, thus, rendering RUS unreliable. It is also found that SVM-SMOTE can be used as a go-to resampling method when paired with the Random Forest and Gaussian Naïve Bayes classifiers, whereby a high F1 score is recorded for all activity classes that are severely and moderately imbalanced. Additionally, the deep learning method called Multilayer Perceptron recorded high F1 scores for all activity classes even when no resampling method was applied. [ABSTRACT FROM AUTHOR]

Simple Summary: To treat diseases caused by genetic mutations, such as mutations in genes and cancer cells, genomic medicine is being promoted to identify disease-causing variants in individual patients using comprehensive genetic analysis (next-generation sequencing, or NGS) for diagnosis and treatment. However, clinical interpretation of the large amount of variant data output by NGS is a time-consuming task and has become a bottleneck in the promotion of genomic medicine. Although AI development to support this task has been conducted in various fields, none has yet been realized that has both high estimation accuracy and explainability at the same time. Therefore, we propose an AI with high estimation accuracy and explanatory power, which will eliminate the bottlenecks in genomic medicine. Background: To treat diseases caused by genetic variants, it is necessary to identify disease-causing variants in patients. However, since there are a large number of disease-causing variants, the application of AI is required. We propose AI to solve this problem and report the results of its application in identifying disease-causing variants. Methods: To assist physicians in their task of identifying disease-causing variants, we propose an explainable AI (XAI) that combines high estimation accuracy with explainability using a knowledge graph. We integrated databases for genomic medicine and constructed a large knowledge graph that was used to achieve the XAI. Results: We compared our XAI with random forests and decision trees. Conclusion: We propose an XAI that uses knowledge graphs for explanation. The proposed method achieves high estimation performance and explainability. This will support the promotion of genomic medicine. [ABSTRACT FROM AUTHOR]

Regenerative agricultural practices are a suitable path to feed the global population. Integrated Crop–livestock systems (ICLSs) are key approaches once the area provides animal and crop production resources. In Brazil, the expectation is to increase the area of ICLS fields by 5 million hectares in the next five years. However, few methods have been tested regarding spatial and temporal scales to map and monitor ICLS fields, and none of these methods use SAR data. Therefore, in this work, we explored the potential of three machine and deep learning algorithms (random forest, long short-term memory, and transformer) to perform early-season (with three-time windows) mapping of ICLS fields. To explore the scalability of the proposed methods, we tested them in two regions with different latitudes, cloud cover rates, field sizes, landscapes, and crop types. Finally, the potential of SAR (Sentinel-1) and optical (Sentinel-2) data was tested. As a result, we found that all proposed algorithms and sensors could correctly map both study sites. For Study Site 1(SS1), we obtained an overall accuracy of 98% using the random forest classifier. For Study Site 2, we obtained an overall accuracy of 99% using the long short-term memory net and the random forest. Further, the early-season experiments were successful for both study sites (with an accuracy higher than 90% for all time windows), and no significant difference in accuracy was found among them. Thus, this study found that it is possible to map ICLSs in the early-season and in different latitudes by using diverse algorithms and sensors. [ABSTRACT FROM AUTHOR]

Airborne hyperspectral data has high spectral-spatial information. However, how to mine and use this information effectively is still a great challenge. Recently, a three-dimensional convolutional neural network (3D-CNN) provides a new effective way of hyperspectral classification. However, its capability of data mining in complex urban areas, especially in cloud shadow areas has not been validated. Therefore, a 3D-1D-CNN model was proposed for feature extraction in complex urban with hyperspectral images affected by cloud shadows. Firstly, spectral composition parameters, vegetation index, and texture characteristics were extracted from hyperspectral data. Secondly, the parameters were fused and segmented into many S × S × B patches which would be input into a 3D-CNN classifier for feature extraction in complex urban areas. Thirdly, Support Vector Machine (SVM), Random Forest (RF),1D-CNN, 3D-CNN, and 3D-2D-CNN classifiers were also carried out for comparison. Finally, a confusion matrix and Kappa coefficient were calculated for accuracy assessment. The overall accuracy of the proposed 3D-1D-CNN is 96.32%, which is 23.96%, 11.02%, 5.22%, and 0.42%, much higher than that of SVM, RF, 1D-CNN, or 3D-CNN, respectively. The results indicated that 3D-1D-CNN could mine spatial-spectral information from hyperspectral data effectively, especially that of grass and highway in cloud shadow areas with missing spectral information. In the future, 3D-1D-CNN could also be used for the extraction of urban green spaces. [ABSTRACT FROM AUTHOR]

Mei et al. [[13]] studied the two-dimensional direction of arrival (2D-DOA) estimation problem in a switching uniform circular array (SUCA), and proposed a covariance matrix completion algorithm for 2D-DOA estimation in a SUCA. The Special Issue "Signal Processing and Machine Learning for Smart Sensing Applications" focused on the publication of advanced signal processing methods by means of state-of-the-art machine learning technologies for smart sensing applications. Monte-Carlo simulations are performed to evaluate the GSCA algorithm, the spatial spectrum plots and RMSE curves demonstrated that the GSCA algorithm can give reasonable results of underdetermined DOA estimation in SCA. The defect detection algorithm based on YOLO algorithms made it possible to obtain results more quickly and with higher mean average precision (mAP) than that of existing methods. [Extracted from the article]

The integration of Micro Electronic Mechanical Systems (MEMS) sensor technology in smartphones has greatly improved the capability for Human Activity Recognition (HAR). By utilizing Machine Learning (ML) techniques and data from these sensors, various human motion activities can be classified. This study performed experiments and compiled a large dataset of nine daily activities, including Laying Down, Stationary, Walking, Brisk Walking, Running, Stairs-Up, Stairs-Down, Squatting, and Cycling. Several ML models, such as Decision Tree Classifier, Random Forest Classifier, K Neighbors Classifier, Multinomial Logistic Regression, Gaussian Naive Bayes, and Support Vector Machine, were trained on sensor data collected from accelerometer, gyroscope, and magnetometer embedded in smartphones and wearable devices. The highest test accuracy of 95% was achieved using the random forest algorithm. Additionally, a custom-built Bidirectional Long-Short-Term Memory (Bi-LSTM) model, a type of Recurrent Neural Network (RNN), was proposed and yielded an improved test accuracy of 98.1%. This approach differs from traditional algorithmic-based human activity detection used in current wearable technologies, resulting in improved accuracy. [ABSTRACT FROM AUTHOR]

Background: Cardiovascular diseases (CVDs) are a leading cause of death worldwide. Deep learning methods have been widely used in the field of medical image analysis and have shown promising results in the diagnosis of CVDs. Methods: Experiments were performed on 12-lead electrocardiogram (ECG) databases collected by Chapman University and Shaoxing People's Hospital. The ECG signal of each lead was converted into a scalogram image and an ECG grayscale image and used to fine-tune the pretrained ResNet-50 model of each lead. The ResNet-50 model was used as a base learner for the stacking ensemble method. Logistic regression, support vector machine, random forest, and XGBoost were used as a meta learner by combining the predictions of the base learner. The study introduced a method called multi-modal stacking ensemble, which involves training a meta learner through a stacking ensemble that combines predictions from two modalities: scalogram images and ECG grayscale images. Results: The multi-modal stacking ensemble with a combination of ResNet-50 and logistic regression achieved an AUC of 0.995, an accuracy of 93.97%, a sensitivity of 0.940, a precision of 0.937, and an F1-score of 0.936, which are higher than those of LSTM, BiLSTM, individual base learners, simple averaging ensemble, and single-modal stacking ensemble methods. Conclusion: The proposed multi-modal stacking ensemble approach showed effectiveness for diagnosing CVDs. [ABSTRACT FROM AUTHOR]

Agriculture is the backbone of any country, and plays a viable role in the total gross domestic product (GDP). Healthy and fruitful crops are of immense importance for a government to fulfill the food requirements of its inhabitants. Because of land diversities, weather conditions, geographical locations, defensive measures against diseases, and natural disasters, monitoring crops with human intervention becomes quite challenging. Conventional crop classification and yield estimation methods are ineffective under unfavorable circumstances. This research exploits modern precision agriculture tools for enhanced remote crop yield estimation, and types classification by proposing a fuzzy hybrid ensembled classification and estimation method using remote sensory data. The architecture enhances the pooled images with fuzzy neighborhood spatial filtering, scaling, flipping, shearing, and zooming. The study identifies the optimal weights of the strongest candidate classifiers for the ensembled classification method adopting the bagging strategy. We augmented the imagery datasets to achieve an unbiased classification between different crop types, including jute, maize, rice, sugarcane, and wheat. Further, we considered flaxseed, lentils, rice, sugarcane, and wheat for yield estimation on publicly available datasets provided by the Food and Agriculture Organization (FAO) of the United Nations and the Word Bank DataBank. The ensemble method outperformed the individual classification methods for crop type classification on an average of 13% and 24% compared to the highest gradient boosting and lowest decision tree methods, respectively. Similarly, we observed that the gradient boosting predictor outperformed the multivariate regressor, random forest, and decision tree regressor, with a comparatively lower mean square error value on yield years 2017 to 2021. Further, the proposed architecture supports embedded devices, where remote devices can adopt a lightweight classification algorithm, such as MobilenetV2. This can significantly reduce the processing time and overhead of a large set of pooled images. [ABSTRACT FROM AUTHOR]

With the advent of the Third computing platform of Social Mobility Analytics and Cloud (SMAC), data is getting generated in huge amounts. This huge amount of data is collected for domain-specific information to process them to get required domain-specific information as in real-time health analytics, financial frauds, real-time automated car driving, vital information of patients undergoing robotic surgery, handling cyber threats etc. This huge data, also known as Big Data, is highly unstructured and imbalanced that is not possible for traditional techniques to handle and process. Advancements in computing power, speedy data storage and convergence of SMAC technologies have also contributed to the swift acceptance of the technology. This led to innovative analytical techniques that are data as well as computation intensive. One such technique is Deep Learning which originated from the artificial neural network and found its use in handling many real-life problems involving multidimensional features. The advantage of Feature Learning or Representational Learning makes Deep Learning a wonderful tool for big data analytics. The previous level of hierarchy transfers the feature learning to the next levels and thus complex features are learned through the learning of simpler features at different levels of abstraction. For efficient learning of these features, tuning of hyper-parameters is a mandatory step. The current work incorporates Grid Search for classification to find the best classifier for the classification of Medicare beneficiaries based on two scenarios. The first Scenario is beneficiaries who are affected by cancer and the Second Scenario is where Medicare beneficiaries are provided Gender wise (being a Female beneficiary). By experimenting using these algorithms at 10-fold cross-validation, the best results were achieved in the sensitivity of 99.17 %, Specificity of 97.68 % and accuracy of 98.8 % with Deep Learning Neural Network with Dropout for First Scenario and achieved the best results in the sensitivity of 82.97 %, Specificity of 68.71 % and accuracy of 75.05 % with Random Forest for Second Scenario. [ABSTRACT FROM AUTHOR]

The largest and most diversified group of organisms are insects. Since insects play a crucial role in many ecosystems, it must be precisely identified for efficient management. However, it is difficult and labor-intensive to identify insect species. Due to advancements in deep learning, computer vision, and sensor technologies, there is rising interest in image-based systems for quick, accurate identification. This study investigates a novel hybrid feature set of shallow features from the tiger beetle dataset, such as texture and wavelet features and high-level features from SqueezeNet. The tiger beetle insect is classified into the cicindelini and collyridini classes with a random forest classifier with 97.65 percent accuracy using this hybrid feature set, which considers the texture and structural characteristics of the insect. Thus the technique provides insight into various features and indicates promising future directions for image-based insect identification and species classification relevant to Computer Science, Agriculture, and Ecology research. [ABSTRACT FROM AUTHOR]

Anomaly-based network intrusion detection systems are highly significant in detecting network attacks. Robust machine learning and deep learning models for identifying network intrusion and attack types are proposed in this paper. Proposed models have experimented with the UNSW-NB15 dataset of 49 features for nine different attack samples. Decision Tree classifier produced the best accuracy of 99.05% compared to ensemble models - Random forest(98.96%), Adaboost(97.87%), and XGBoost(98.08%). K-Nearest Neighbour classifier trained for various values of K and best performance obtained for K=7 with the accuracy of 95.58%. A Deep Learning model with two dense layers with ReLU activation and a third dense layer with a Sigmoid activation function is designed for binary classification and produced good accuracy of 98.44% with ADAM optimizer, 80:20 Train-Test Split Ratio. Network attack exploits are detected with an accuracy 95% by XGBoost, Fuzzers attack with 90% accuracy by Random Forest, Generic attacks with 99% accuracy by Random Forest, and Reconnaissance attacks with 79% by Decision Trees. All features are relevant and strong in network attack detection, which eliminates the requirement of feature selection. [ABSTRACT FROM AUTHOR]

Seismic activity prediction has been a challenging research domain: in this regard, accurate prediction using historical data is an intricate task. Numerous machine learning and traditional approaches have been presented lately for seismic activity prediction; however, no generalizable model exists. In this work, we consider seismic activity predication as a binary classification problem, and propose a deep neural network architecture for the classification problem, using historical data from Chile, Hindukush, and Southern California. After obtaining the data for the three regions, a data cleaning process was used, which was followed by a feature engineering step, to create multiple new features based on various seismic laws. Afterwards, the proposed model was trained on the data, for improved prediction of the seismic activity. The performance of the proposed model was evaluated and compared with extant techniques, such as random forest, support vector machine, and logistic regression. The proposed model achieved accuracy scores of 98.28%, 95.13%, and 99.29% on the Chile, Hindukush, and Southern California datasets, respectively, which were higher than the current benchmark model and classifiers. In addition, we also conducted out-sample testing, where the evaluation metrics confirmed the generality of our proposed approach. [ABSTRACT FROM AUTHOR]

Because of the complexity, nonlinearity, and volatility, stock market forecasting is either highly difficult or yields very unsatisfactory outcomes when utilizing traditional time series or machine learning techniques. To cope with this problem and improve the complex stock market's prediction accuracy, we propose a new hybrid novel method that is based on a new version of EMD and a deep learning technique known as long-short memory (LSTM) network. The forecasting precision of the proposed hybrid ensemble method is evaluated using the KSE-100 index of the Pakistan Stock Exchange. Using a new version of EMD that uses the Akima spline interpolation technique instead of cubic spline interpolation, the noisy stock data are first divided into multiple components technically known as intrinsic mode functions (IMFs) varying from high to low frequency and a single monotone residue. The highly correlated sub-components are then used to build the LSTM network. By comparing the proposed hybrid model with a single LSTM and other ensemble models such as the support vector machine (SVM), Random Forest, and Decision Tree, its prediction performance is thoroughly evaluated. Three alternative statistical metrics, namely root means square error (RMSE), mean absolute error (MAE) and mean absolute percentage error (MAPE), are used to compare the aforementioned techniques. The empirical results show that the suggested hybrid Akima-EMD-LSTM model beats all other models taken into consideration for this study and is therefore recommended as an effective model for the prediction of non-stationary and nonlinear complex financial time series data. [ABSTRACT FROM AUTHOR]

Student engagement is a flexible, complicated concept that includes behavioural, emotional, and cognitive involvement. In order for the instructor to understand how the student interacts with the various activities in the classroom, it is essential to predict their participation. The current work aims to identify the best algorithm for predicting student engagement in the classroom. In this paper, we gathered data from VLE and prepared them using a variety of data preprocessing techniques, including the elimination of missing values, normalization, encoding, and identification of outliers. On our data, we ran a number of machine learning (ML) classification algorithms, and we assessed each one using cross-validation methods and many helpful indicators. The performance of the model is evaluated with metrics like accuracy, precision, recall, and AUC scores. The results show that the CATBoost model is having higher accuracy than the rest. This proposed model outperformed in all the aspects compared to previous research. The results part of this paper indicates that the CATBoost model had an accuracy of approximately 92.23%, a precision of 94.40%, a recall of 100%, and an AUC score of 0.9624. The XGBoost predictive model, the random forest model, and the multilayer perceptron model all demonstrated approximately the same performance overall. We compared the AISAR model with Our model achieved an accuracy of 94.64% compared with AISAR 91% model and it concludes that our results are better. The AISAR model had only around 50% recall compared to our models, which had around 92%. This shows that our models return more relevant results, i.e., if our models predict that a student has high engagement, they are correct 94.64% of the time. [ABSTRACT FROM AUTHOR]

Accurate prediction of postoperative complications can inform shared decisions regarding prognosis, preoperative risk-reduction, and postoperative resource use. We hypothesized that multi-task deep learning models would outperform conventional machine learning models in predicting postoperative complications, and that integrating high-resolution intraoperative physiological time series would result in more granular and personalized health representations that would improve prognostication compared to preoperative predictions. In a longitudinal cohort study of 56,242 patients undergoing 67,481 inpatient surgical procedures at a university medical center, we compared deep learning models with random forests and XGBoost for predicting nine common postoperative complications using preoperative, intraoperative, and perioperative patient data. Our study indicated several significant results across experimental settings that suggest the utility of deep learning for capturing more precise representations of patient health for augmented surgical decision support. Multi-task learning improved efficiency by reducing computational resources without compromising predictive performance. Integrated gradients interpretability mechanisms identified potentially modifiable risk factors for each complication. Monte Carlo dropout methods provided a quantitative measure of prediction uncertainty that has the potential to enhance clinical trust. Multi-task learning, interpretability mechanisms, and uncertainty metrics demonstrated potential to facilitate effective clinical implementation. [ABSTRACT FROM AUTHOR]

Applying computer vision and machine learning techniques into sport tests is an effective way to realize "intelligent sports." Facing practical application, we design a real-time and lightweight deep learning network to realize intelligent pull-ups test in this study. The main contributions are as follows: (1) a new self-produced pull-ups dataset is established under the requirement of including a human body and horizontal bar. In addition, a semiautomatic annotating software is developed to enhance annotation efficiency and increase labeling accuracy. (2) A novel lightweight deep network named PEPoseNet is designed to estimate and analyze a human pose in real time. The backbone of the network is made up of the heatmap network and key point network, which conduct human pose estimation based on the key points extracted from a human body and horizontal bar. The depth-wise separable convolution is adopted to speed up the training and convergence. (3) An evaluation criterion of intelligent pull-ups test is defined based on action quality assessment (AQA). The action quality of five states, i.e., ready or end, hang, pull, achieved, and resume in one pull-ups test cycle is automatically graded using a random forest classifier. A mobile application is developed to realize intelligent pull-ups test in real time. The performance of the proposed model and software is confirmed by verification and ablation experiments. The experimental results demonstrated that the proposed PEPoseNet has competitive performance to the state of the art. Its PCK @ 0.2 and frames per second (FPS) achieved were 83.8 and 30 fps, respectively. The mobile application has promising application prospects in pull-ups test under complex scenarios. [ABSTRACT FROM AUTHOR]

Diabetes mellitus presents a high prevalence around the world. A common and long-term derived complication is diabetic foot ulcers (DFUs), which have a global prevalence of roughly 6.3%, and a lifetime incidence of up to 34%. Infrared thermograms, covering the entire plantar aspect of both feet, can be employed to monitor the risk of developing a foot ulcer, because diabetic patients exhibit an abnormal pattern that may indicate a foot disorder. In this study, the publicly available INAOE dataset composed of thermogram images of healthy and diabetic subjects was employed to extract relevant features aiming to establish a set of state-of-the-art features that efficiently classify DFU. This database was extended and balanced by fusing it with private local thermograms from healthy volunteers and generating synthetic data via synthetic minority oversampling technique (SMOTE). State-of-the-art features were extracted using two classical approaches, LASSO and random forest, as well as two variational deep learning (DL)-based ones: concrete and variational dropout. Then, the most relevant features were detected and ranked. Subsequently, the extracted features were employed to classify subjects at risk of developing an ulcer using as reference a support vector machine (SVM) classifier with a fixed hyperparameter configuration to evaluate the robustness of the selected features. The new set of features extracted considerably differed from those currently considered state-of-the-art but provided a fair performance. Among the implemented extraction approaches, the variational DL ones, particularly the concrete dropout, performed the best, reporting an F1 score of 90% using the aforementioned SVM classifier. In comparison with features previously considered as the state-of-the-art, approximately 15% better performance was achieved for classification. [ABSTRACT FROM AUTHOR]

Data in the form of images are now generated at an unprecedented rate. A case in point is remote sensing images (RSI), now available in large-scale RSI archives, which have attracted a considerable amount of research on image classification within the remote sensing community. The basic task of single-target multi-class image classification considers the case where each image is assigned exactly one label from a predefined finite set of class labels. Recently, however, image annotations have become increasingly complex, with images labeled with several labels (instead of just one). In other words, the goal is to assign multiple semantic categories to an image, based on its high-level context. The corresponding machine learning tasks is called multi-label classification (MLC). The classification of RSI is currently predominantly addressed by deep neural network (DNN) approaches, especially convolutional neural networks (CNNs), which can be utilized as feature extractors as well as end-to-end methods. After only considering single-target classification for a long period, DNNs have recently emerged that address the task of MLC. On the other hand, trees and tree ensembles for MLC have a long tradition and are the best-performing class of MLC methods, but need predefined feature representations to operate on. In this work, we explore different strategies for model training based on the transfer learning paradigm, where we utilize different families of (pre-trained) CNN architectures, such as VGG, EfficientNet, and ResNet. The architectures are trained in an end-to-end manner and used in two different modes of operation, namely, as standalone models that directly perform the MLC task, and as feature extractors. In the latter case, the learned representations are used with tree ensemble methods for MLC, such as random forests and extremely randomized trees. We conduct an extensive experimental analysis of methods over several publicly available RSI datasets and evaluate their effectiveness in terms of standard MLC measures. Of these, ranking-based evaluation measures are most relevant, especially ranking loss. The results show that, for addressing the RSI-MLC task, it is favorable to use lightweight network architectures, such as EfficientNet-B2, which is the best performing end-to-end approach, as well as a feature extractor. Furthermore, in the datasets with a limited number of images, using traditional tree ensembles for MLC can yield better performance compared to end-to-end deep approaches. [ABSTRACT FROM AUTHOR]

The invasive pest Dendroctonus valens has spread to northeast China, causing serious economic and ecological losses. Early detection and disposal of infested trees is critical to prevent its outbreaks. This study aimed to evaluate the potential of an unmanned aerial vehicle (UAV)-based hyperspectral image for early detection of D. valens infestation at the individual tree level. We compared the spectral characteristics of Pinus tabuliformis in three states (healthy, infested and dead), and established classification models using three groups of features (reflectance, derivatives and spectral vegetation indices) and two algorithms (random forest and convolutional neural network). The spectral features of dead trees were clearly distinct from those of the other two classes, and all models identified them accurately. The spectral changes of infested trees occurred mainly in the visible region, but it was difficult to distinguish infested from healthy trees using random forest classification models based on reflectance and derivatives. The random forest model using spectral vegetation indices and the convolutional neural network model performed better, with an overall accuracy greater than 80% and a recall rate of infested trees reaching 70%. Our results demonstrated the great potential of hyperspectral imaging and deep learning for the early detection of D. valens infestation. The convolutional neural network proposed in this study can provide a reference for the automatic detection of early D. valens infestation using UAV-based multispectral or hyperspectral images in the future. [ABSTRACT FROM AUTHOR]

Deep learning has drawn great attention in the financial field due to its powerful ability in nonlinear fitting, especially in the studies of asset pricing. In this paper, we proposed a long short-term memory option pricing model with realized skewness by fully considering the asymmetry of asset return in emerging markets. It was applied to price the ETF50 options of China. In order to emphasize the improvement of this model, a comparison with a parametric method, such as Black-Scholes (BS), and machine learning methods, such as support vector machine (SVM), random forests and recurrent neural network (RNN), was conducted. Moreover, we also took the characteristic of heavy tail into consideration and studied the effect of realized kurtosis on pricing to prove the robustness of the skewness. The empirical results indicate that realized skewness significantly improves the pricing performance of LSTM among moneyness states except for in-the-money call options. Specifically, the LSTM model with realized skewness outperforms the classical method and other machine learning methods in all metrics. [ABSTRACT FROM AUTHOR]

Benefiting from the high cadence and spatial resolution of the new generation of Earth observation satellites, remote sensing technology is allowing us to derive more valuable information for the agricultural sector. Crop classification is one of the fundamental information derivatives from Earth observation data researchers used for food security, crop monitoring, and economic assessment. The robustness of a crop classification model to variations in environmental and management conditions due to time and location is one of the crucial requirements. To achieve this, we developed a novel self-supervised method using the advantage of unlabeled samples and transformer architectures. We used six different areas in Germany and four years to evaluate the robustness of the model. Our experiments showed that self-supervised deep learning methods could provide a significant advantage in handling these variations. In some cases, we observed around 30 percentage points improvements in F1-score performance compared to a Random Forest based model. [ABSTRACT FROM AUTHOR]

Membrane-based cells are the fundamental structural and functional units of organisms, while evidences demonstrate that liquid–liquid phase separation (LLPS) is associated with the formation of membraneless organelles, such as P-bodies, nucleoli and stress granules. Many studies have been undertaken to explore the functions of protein phase separation (PS), but these studies lacked an effective tool to identify the sequence segments that critical for LLPS. In this study, we presented a novel software called dSCOPE (http://dscope.omicsbio.info) to predict the PS-driving regions. To develop the predictor, we curated experimentally identified sequence segments that can drive LLPS from published literature. Then sliding sequence window based physiological, biochemical, structural and coding features were integrated by random forest algorithm to perform prediction. Through rigorous evaluation, dSCOPE was demonstrated to achieve satisfactory performance. Furthermore, large-scale analysis of human proteome based on dSCOPE showed that the predicted PS-driving regions enriched various protein post-translational modifications and cancer mutations, and the proteins which contain predicted PS-driving regions enriched critical cellular signaling pathways. Taken together, dSCOPE precisely predicted the protein sequence segments critical for LLPS, with various helpful information visualized in the webserver to facilitate LLPS-related research. [ABSTRACT FROM AUTHOR]

MicroRNA (miRNA)–target interaction (MTI) plays a substantial role in various cell activities, molecular regulations and physiological processes. Published biomedical literature is the carrier of high-confidence MTI knowledge. However, digging out this knowledge in an efficient manner from large-scale published articles remains challenging. To address this issue, we were motivated to construct a deep learning-based model. We applied the pre-trained language models to biomedical text to obtain the representation, and subsequently fed them into a deep neural network with gate mechanism layers and a fully connected layer for the extraction of MTI information sentences. Performances of the proposed models were evaluated using two datasets constructed on the basis of text data obtained from miRTarBase. The validation and test results revealed that incorporating both PubMedBERT and SciBERT for sentence level encoding with the long short-term memory (LSTM)-based deep neural network can yield an outstanding performance, with both F1 and accuracy being higher than 80% on validation data and test data. Additionally, the proposed deep learning method outperformed the following machine learning methods: random forest, support vector machine, logistic regression and bidirectional LSTM. This work would greatly facilitate studies on MTI analysis and regulations. It is anticipated that this work can assist in large-scale screening of miRNAs, thereby revealing their functional roles in various diseases, which is important for the development of highly specific drugs with fewer side effects. Source code and corpus are publicly available at https://github.com/qi29. [ABSTRACT FROM AUTHOR]

This paper aims to quantify the errors in the provided agricultural crop types, estimate the possible error rate in the available dataset, and propose a correction strategy. This quantification could establish a confidence criterion useful for decisions taken on this data or to have a better apprehension of the possible consequences of using this data in learning downstream functions such as classification. We consider two agricultural label errors: crop type mislabels and mis-split crops. To process and correct these errors, we design a two-step methodology. Using class-specific convolutional autoencoders applied to synthetic aperture radar (SAR) time series of free-to-use and temporally dense Sentinel-1 data, we detect out-of-distribution temporal profiles of crop time series, which we categorize as one out of the three following possibilities: crop edge confusion, incorrectly split crop areas, and potentially mislabeled crop. We then relabel crops flagged as mislabeled using an Otsu threshold-derived confidence criteria. We numerically validate our methodology using a controlled disruption of labels over crops of confidence. We then compare our methods to supervised algorithms and show improved quality of relabels, with up to 98% correct relabels for our method, against up to 91% for Random Forest-based approaches. We show a drastic decrease in the performance of supervised algorithms under critical conditions (smaller and larger amounts of introduced label errors), with Random Forest falling to 56% of correct relabels against 95% for our approach. We also explicit the trade-off made in the design of our method between the number of relabels, and their quality. In addition, we apply this methodology to a set of agricultural labels containing probable mislabels. We also validate the quality of the corrections using optical imagery, which helps highlight incorrectly cut crops and potential mislabels. We then assess the applicability of the proposed method in various contexts and scales and present how it is suitable for verifying and correcting farmers' crop declarations. [ABSTRACT FROM AUTHOR]

Simple Summary: With the increased development of pig farming intensification, air quality and odor emissions in pig houses are gradually attracting attention. Among them, ammonia is considered to be an important environmental indicator of pig house. Excessive accumulation of ammonia can seriously affect the growth status of pigs and also cause a potential health risk to farm workers. Therefore, it is very important to recognize the changes of ammonia in pig houses and to discharge ammonia in time for the welfare farming of pigs. In this study, three traditional machine learning algorithms and three deep learning algorithms were selected to predict the ammonia concentration in a pig house. Based on them, important environmental parameters and promising algorithms were screened out and the algorithms were evaluated for optimization. The results of the study can provide a reference for air quality regulation in pig houses. Accurately predicting the air quality in a piggery and taking control measures in advance are important issues for pig farm production and local environmental management. In this experiment, the NH3 concentration in a semi-automatic piggery was studied. First, the random forest algorithm (RF) and Pearson correlation analysis were combined to analyze the environmental parameters, and nine input schemes for the model feature parameters were identified. Three kinds of deep learning and three kinds of conventional machine learning algorithms were applied to the prediction of NH3 in the piggery. Through comparative experiments, appropriate environmental parameters (CO2, H2O, P, and outdoor temperature) and superior algorithms (LSTM and RNN) were selected. On this basis, the PSO algorithm was used to optimize the hyperparameters of the algorithms, and their prediction performance was also evaluated. The results showed that the R2 values of PSO-LSTM and PSO-RNN were 0.9487 and 0.9458, respectively. These models had good accuracy when predicting NH3 concentration in the piggery 0.5 h, 1 h, 1.5 h, and 2 h in advance. This study can provide a reference for the prediction of air concentrations in pig house environments. [ABSTRACT FROM AUTHOR]

This study, which aims to make predictions using a previously unused deep hybrid Convolutional Neural Network-Support Vector Regression approach for hydropower generation, was carried out using a unique meteorological data set as input data. The unique dataset was collected from Kaman Meteorology Directorate and Hirfanlı HEPP in Turkey, from 2007 to 2021, to estimate the HEPP's Net Head and Hydroelectric Power Generation on a daily basis. The performances of the prediction models were benchmarked in addition to the used CNN-SVR model with Machine Learning (ML) models (Boosting Random Forest Regression (BRFR) and Weighted K-Nearest Neighbor Regression (WKNNR)) and Deep Learning (DL) models (Long-Short Term Memory (LSTM) and Deep Belief Network (DBN)). The comparison results of the used CNN-SVR model with other alternative models showed that CNN-SVR performed effectively with the highest correlation coefficient of 0.971 for NH and 0.968 for PP. [ABSTRACT FROM AUTHOR]

This study proposed a comprehensive approach that utilized PlanetScope imagery for classifying tropical-marine benthic habitats after retrieving bathymetry from ICESat-2 data and water-column correction for areas around Lyson Islands, Vietnam. Exact bathymetry derivation and water column correction were applied to the PlanetScope imagery, making it an effective method for mapping marine benthic habitats. Water column correction was achieved by applying Depth Invariant Index (DII) and Bottom Reflectance Index (BRI). Moreover, two conventional machine learning algorithms, including Random Forest and Support Vector Machine, and a current deep Convolutional Neural Network (CNN) was employed to classify the benthic features. The overall accuracy of these classifiers are 80.74%, 84.19%, and 89.80% with the BRI, 80.17%, 82.75%, and 87.85% with the DII compared to 37.64%, 42.5%, and 47.2% of without corrected water columns respectively. The CNN model demonstrated that the approach significantly maximizes the improvement in benthic classification results in coastal region. [ABSTRACT FROM AUTHOR]

The estimation of an image geo-site solely based on its contents is a promising task. Compelling image labelling relies heavily on contextual information, which is not as simple as recognizing a single object in an image. An Auto-Encode-based support vector machine approach is proposed in this work to estimate the image geo-site to address the issue of misclassifying the estimations. The proposed method for geo-site estimation is conducted using a dataset consisting of 125 classes of various images captured within 125 countries. The proposed work uses a convolutional Auto-Encode for training and dimensionality reduction. After that, the acquired preprocessed input dataset is further processed by a multi-label support vector machine. The performance assessment of the proposed approach has been accomplished using accuracy, sensitivity, specificity, and F1-score as evaluation parameters. Eventually, the proposed approach for image geo-site estimation presented in this article outperforms Auto-Encode-based K-Nearest Neighbor and Auto-Encode-Random Forest methods. [ABSTRACT FROM AUTHOR]