Search

• Hybrid models successfully predicted key performance metrics for antibody production in mammalian cells. • Such models allow predictions under varying process conditions, without the need to re-estimate parameters. • Performance could be predicted for a different product under similar process conditions with reasonable accuracy. • Algorithm choice did not significantly impact model performance, but computational requirements differed. Among the many approaches to build predictive models for bioprocesses, hybrid models have gained increased attention due to their ability to combine data driven approaches with process knowledge. This study develops and compares hybrid models that predict the performance of a mammalian CHO cell culture producing a therapeutic product. Three machine learning algorithms, MLP, Random Forest and XGBoost regressors are compared to understand the effect of algorithm choice on the rates predicted and overall performance of the hybrid model. When combined in series with mechanistic equations, all three algorithms could predict next day Viable Cell Density, titer and the cumulative amount of glucose consumed with low error, although decision tree-based algorithm required less computational time for training. The models performed well using a dataset for a new product, without further model training. Retraining the model on a subset of data from the new product improved the prediction accuracy of VCD and titer. The potential for application and further model developments are also outlined. [Display omitted] [ABSTRACT FROM AUTHOR]

• Relationships among amyloid PET, MRI, and cognition are unclear in Alzheimer. • Deep-SHAP is proposed to investigate regional imaging vs. cognition relationships. • Deep-SHAP identified the modality-specific brain regions relevant to cognition. Mild cognitive impairment (MCI)/Alzheimer's disease (AD) is associated with cognitive decline beyond normal aging and linked to the alterations of brain volume quantified by magnetic resonance imaging (MRI) and amyloid-beta (Aβ) quantified by positron emission tomography (PET). Yet, the complex relationships between these regional imaging measures and cognition in MCI/AD remain unclear. Explainable artificial intelligence (AI) may uncover such relationships. We integrate the AI-based deep learning neural network and Shapley additive explanations (SHAP) approaches and introduce the Deep-SHAP method to investigate the multivariate relationships between regional imaging measures and cognition. After validating this approach on simulated data, we apply it to real experimental data from MCI/AD patients. Deep-SHAP significantly predicted cognition using simulated regional features and identified the ground-truth simulated regions as the most significant multivariate predictors. When applied to experimental MRI data, Deep-SHAP revealed that the insula, lateral occipital, medial frontal, temporal pole, and occipital fusiform gyrus are the primary contributors to global cognitive decline in MCI/AD. Furthermore, when applied to experimental amyloid Pittsburgh compound B (PiB)-PET data, Deep-SHAP identified the key brain regions for global cognitive decline in MCI/AD as the inferior temporal, parahippocampal, inferior frontal, supratemporal, and lateral frontal gray matter. Deep-SHAP method uncovered the multivariate relationships between regional brain features and cognition, offering insights into the most critical modality-specific brain regions involved in MCI/AD mechanisms. [ABSTRACT FROM AUTHOR]

Landslides pose a significant threat to lives and socio-economic stability globally. In this study, we conducted a comprehensive landslide susceptibility mapping (LSM) in the western region of Cameroon, focusing on Bafoussam and its surroundings. The integration of multi-criteria decision analysis models (AHP), statistical methods (Information Value IV, Shannon Entropy SE, Frequency Ratio FR), and machine learning algorithms (Naïve Bayes and Logistic Regression) provided a robust assessment of landslide risk. Our analysis, based on 54 recorded landslides, carefully selected Landslide Conditioning Factors (LCF), and influential parameters such as lithology, slope, altitude, and precipitation, resulted in susceptibility maps categorizing the area into five risk zones. The Spatial distribution shows the centre and northwestern regions as high-risk areas. Model sensitivity differences underscore the need for tailored LSM selection. Validation using the Area Under Curve/Receiver Operating Characteristics (AUC/ROC) method indicates the LR and NB methods have the highest accuracy (82.7% and 84.1%, respectively). Comparative analysis of landslide events in Gouaché, Sichuan, Souk Ahras, and Kekem reveals correlations between heavy rainfall and geological conditions. The study supplies valuable insights for decision-makers in landslide-prone areas, emphasizing the importance of integrating multiple methodologies for comprehensive risk assessment and management. [Display omitted] • A comprehensive landslide susceptibility mapping (LSM) was conducted in the west region of Cameroon, focusing on Bafoussam and its surroundings. • The study utilized a combination of Multi-Criteria Decision Analysis models, statistical methods, and machine learning algorithms for a robust assessment of landslide risk. • Susceptibility maps classified the area into five risk zones, with the centre and northwestern regions identified as high-risk areas. • Validation using the AUC/ROC method showed the Logistic Regression and Naïve Bayes methods had the highest accuracy (82.7% and 84.1%, respectively). • The study underscores the need for a tailored approach to LSM selection and the importance of integrating multiple methodologies for comprehensive risk assessment and management. [ABSTRACT FROM AUTHOR]

Diabetic vascular complications (DVC) are the main cause of death in diabetic patients. However, there is a lack of effective biomarkers or convenient methods for early diagnosis of DVC. In this study, the salivary glycopatterns from 130 of healthy volunteers (HV), 139 patients with type 2 diabetes mellitus (T2DM) and 167 patients with DVC were case-by-case analyzed by using lectin microarrays. Subsequently, diagnostic models were developed using logistic regression and machine learning algorithms based on the data of lectin microarrays in training set. The performance of diagnostic models was evaluated in an independent blind cohort. The results of lectin microarrays indicated that the glycopatterns identified by 16 lectins (e.g. BS-I, PWM and EEL) were significantly altered in DVC patients compared with patients with T2DM, which suggested the alterations in salivary glycopatterns could reflect onset of DVC. Notably, K-Nearest Neighbor (KNN) model exhibited better performance for distinguishing DVC (accuracy: 0.939) than other models in blind cohort. The integrated classifier, which combined three machine learning models, exhibited a higher overall accuracy (≥ 0.933) than other models in blind cohort. Our study provided a cost-effective and non-invasive method for auxiliary diagnosis DVC based on the combination of salivary glycopatterns and machine learning algorithms. • There were no convenient methods for early diagnosis of diabetic vascular complications (DVC) • Salivary glycopatterns showed significantly altered in DVC patients compared to type 2 diabetes mellitus patients • The K-Nearest Neighbor (KNN) model exhibited better performance for distinguishing DVC (accuracy: 0.939) than other models in blind cohort • The integrated classifier exhibited a higher overall accuracy (≥ 0.933) than other models in blind cohort [ABSTRACT FROM AUTHOR]

Even though the World Health Organization declared the end of the pandemic, COVID-19 is still considered an endemic disease, affecting many people worldwide. As a result, hospitalization rates can continue to strain healthcare systems, particularly in low and middle-income countries. Thus, this article presents two predictive Machine Learning (ML) models to identify COVID-19 patients more likely to be hospitalized or face a higher mortality risk, aiming to assist physicians and healthcare administrators in making informed decisions. Previous research primarily focused on using expensive input data, such as clinical examinations and images, to implement ML models. In contrast, we implemented the models using information on patients' demographics, vaccination history, symptoms, and underlying health conditions. We trained the ML algorithms using data from the OpenDataSus database, focusing on patients undergoing hospital treatment in São Paulo, Brazil. Our study tested 14 different ML algorithms to determine their effectiveness: gradient boosting, light gradient boosting machine, logistic regression, linear discriminant analysis, ridge, AdaBoost, random forest, support vector machine, k-nearest neighbor, extra tree, decision tree, naive Bayes, quadratic discriminant analysis, and dummy. The results showed that the gradient-boosting model outperformed the others, achieving an accuracy rate of 83% and an Area Under the Curve (AUC) of 0.89 for predicting mortality risk. Moreover, the model achieved an accuracy rate of 71% and an AUC of 0.75 for predicting hospitalization risk. Our analysis evidence that including information on the number of vaccine doses can enhance the prediction results for hospitalization and mortality of patients. Based on these significant findings, we have developed a web application to assist healthcare professionals in identifying patients at a higher risk of severe outcomes from COVID-19. • Models for identifying COVID-19 patients at risk of mortality and hospitalization. • Gathered São Paulo hospital data for COVID-19 model training. • Evaluated 14 machine Learning algorithms; gradient boosting proved most effective. • Vaccine doses can improve the risk of mortality and hospitalization. • Web app aids healthcare professionals in identifying high-risk COVID-19 cases. [ABSTRACT FROM AUTHOR]

Soft 2D tactile sensors are becoming increasingly important in robotics and human-machine interaction. In this paper, we propose a new approach to develop a soft tactile sensor using fiber Bragg gratings (FBGs) and machine learning algorithms. The sensor consists of a layer of silicone elastomer with embedded 192 FBGs that can detect deformations caused by point impact. The FBG responses are then processed by machine learning algorithms to measure the position and the force of impacts with the mean absolute errors of 2.1 mm and 0.34 N, respectively. [Display omitted] • 2D flexible tactile sensor with array of 192 FBG sensors was shown. • Experimental calibration of 2D tactile sensors with assistance of machine learning algorithm was presented. • Prediction of pressing force with mean absolute error of 0.34 N was demonstrated. • Prediction of position of the impact with mean absolute error of 2.1 mm was achieved. [ABSTRACT FROM AUTHOR]

Climate changes and global warming increase risk of recurrent extreme and complex climatic features. It necessitates accurate modeling and forecasting of climate phenomena for sustainable development goals. However, machine learning algorithms and advanced statistical models are extensively employed to analyze complex data and make predictions related to climate phenomena. It is important to have comprehensive knowledge to use these models and consider their potential implications. This study aims to evaluate and compare some popular machine learning and probabilistic methods by analyzing various time series indices associated with precipitation and temperature. For application, time series data of Standardized Precipitation Temperature Index (SPTI), Standardized Temperature Index (STI), Standardized Compound Drought and Heat Index (SCDHI), and Biased Diminished Weighted Regional Drought Index (BDWRDI) are used from various meteorological regions of Pakistan. The performance of each algorithm is compared using Residual Mean Square Error (RMSE) and Mean Average Error (MAE). The outcomes associated with this research indicate a higher preference of neural networks over machine learning methods in the training sets. However, the efficiency varies from model to model, indicator to indicator, time scale to time scale, and location to location during the testing phase. The most appropriate models are found by considering a list of candidates forecasting models and investigating the performance of each model. [ABSTRACT FROM AUTHOR]

Blockchain technology, with its peer-to-peer trading feature, influences the management of energy consumption by offering the potential to transform transparency, efficiency, and sustainability within the energy sector. Nonetheless, there is a need to develop analytical decision-making models tailored for managing peer-to-peer energy transactions to improve energy resilience. Therefore, the purpose of this paper is to address the research question: How can energy distribution systems be protected via blockchain technology to enhance energy resilience and mitigate vulnerabilities to disruptions ? This paper employs a conceptual research model design and a mathematical decision-making model to address the research question by capturing the peer-to-peer trading capability of blockchain technology. The theory of planned behavior provides theoretical explanations for the proposed model. The sample includes longitudinal energy consumption data from 2015 to 2023 in Texas. The findings indicate a significant improvement in energy efficiency along with a considerable decrease in total electricity consumption. Post hoc analysis results reveal that the seasonal autoregressive integrated moving average algorithm is effective as a reliable input for the proposed mathematical model. The significant implications are to implement blockchain-based smart grids in which energy systems become more resilient to disruptions, as the peer-to-peer capability enables users to trade energy. The proposed model suggests that energy will be used more efficiently and effectively. This paper contributes to prior works by introducing a mathematical model that captures the trading behavior of energy consumers. Moreover, this paper proposes the SARIMA algorithm to predict energy demand. • A mathematical peer-to-peer trading model is presented to improve resilience. • A blockchain-based smart grid is presented for consumer energy consumption. • Five machine learning algorithms are presented in energy demand management. • Longitudinal energy consumption data of Texas is used in the model. • Theoretical support is provided using the theory of planned behavior. [ABSTRACT FROM AUTHOR]

[Display omitted] • A dual-scale non-invasive technique for crack detection is demonstrated. • Resonance-based device is evaluated for its attenuation and frequency shift. • A comparison between sensors operating in microwave and radiowave regime is evaluated. • The microwave regime device show superior quality factor with ∼250 % improved sensitivity. • Out of multiple machine learning algorithms, k-NN algorithm proved to be most accurate with accuracy of around ∼99 %. Condition monitoring of structures built using metallic supports is a significant issue in structural engineering. Through this work, design, fabrication, and optimization of a dual-scale resonance-based microwave sensor is discussed for monitoring cracks developed over metallic surfaces. These sensors were initially simulated and later fabricated in micro-wave (4.59 GHz) and radio-wave region (418 MHz) for which the power attenuation and frequency shifts were observed. With the structure having crack depth of 2 mm, microwave sensor showed the power shift ∼24 %, while the radio-wave range sensors showed ∼9 %. The corresponding frequency shifts were 2.4 % and 0.85 %, respectively. The microwave sensor (power shift data) showed ∼280 % more sensitivity than the radio-wave sensor. By using dual scale, comprising of both frequency and power shifts, the microwave sensor classification was further improved to 99.72 %, as compared to their individual counterparts. A non-destructive crack detection system in microwave regime is hence demonstrated, which can be further explored for range of crack studies in metallic structures. [ABSTRACT FROM AUTHOR]

A soft fiber-reinforced bending actuator (SFRBA) is a multi-material system that plays a crucial role in robotics applications due to its high output force and robust bending motion. However, the multi-material composition of SFRBAs causes significant structural nonlinearity. This nonlinear behavior is difficult to capture using traditional analytical models. This study presents a method for modeling the SFRBA using finite element method (FEM) and nonlinear machine learning algorithms (MLAs). First, the key structural parameters of the SFRBA are defined and selected as input variables for the method. An accurate FEM, considering varying actuation pressure, is then employed to generate a simulation training dataset. Subsequently, three nonlinear MLAs, including polynomial regression (PLR), extreme gradient boosting regression (XGBoostR), and normalized multilayer perceptron regression (NMLPR), are developed to model the bending angle of the SFRBA. Moreover, transfer learning is deployed to improve the accuracy and convergence speed of the optimal NMLPR. Experimental measurements are conducted to validate the established MLAs, and the results demonstrate that the refined NMLPR outperforms the other models, yielding an average Root Mean Square Error (RMSE) of 7.935° and Mean Absolute Percentage Error (MAPE) of 6.45%. Furthermore, the refined NMLPR is implemented in a feedback control loop to showcase its real-time ability to convert actuation pressure information into bending angles. The results exhibit excellent control performance, with an average MAPE of less than 6% and a low time delay of 0.0875 s. This work exemplifies a methodology that combines FEM and MLAs for modeling SFRBAs, paving the way for their further development and practical application. [Display omitted] • Multiple compositions of the SFRBAs increase the structural nonlinearity. • FEM-based MLAs are proposed for modeling the bending angle of SFRBA. • The result of optimal MLA yields an average RMSE of 7.935°. • The best MLA is used in a feedback control loop to showcase its real-time ability. [ABSTRACT FROM AUTHOR]

In recent years, there has been a notable surge in the prevalence of cardiovascular diseases (CVD), presenting a significant global public health challenge and a leading cause of mortality worldwide. Among the myriad complications stemming from CVD, heart failure stands out as a critical concern. Addressing heart failure through surgical means poses considerable challenges. The primary objective of this research is to identify pivotal attributes linked to heart failure and employ diverse machine learning methodologies to predict its occurrence, thereby enabling early estimation of mortality rates associated with heart failure. Leveraging a heart failure dataset, we conducted comprehensive model construction using pre-processing techniques such as feature scaling and correlation analysis. The Extreme Gradient Boosting (XGBoost) method was instrumental in evaluating feature relevance, leading to the selection of two distinct datasets: the whole dataset and the XGBoost dataset. In conclusion, we employed thirteen machine learning methods to predict the occurrence of death events within these datasets. Fine-tuning hyperparameters significantly enhanced model performance. Notably, our model demonstrated exceptional performance on this dataset, achieving the highest accuracy of 85.23% with Random Forest on the whole dataset and 86.36% with Flexible Discriminant Analysis on the XGBoost dataset. [Display omitted] • This research addresses the rising global prevalence of cardiovascular diseases, focussing heart failure. • The study aims to predict heart failure occurrence, enabling early estimation of mortalityrates. • Employing model, feature scaling and correlation analysis method to identify pivotal attributes linked to heart failure. [ABSTRACT FROM AUTHOR]

The urban thermal environment undergoes significant influences from changes in land use/land cover (LULC). This article uses CA-ANN and ANN algorithms to forecast LULC and changes in the urban thermal environment in Nanjing for the years 2030 and 2040. It investigates the interplay between LULC changes, land surface temperature (LST), and the urban thermal field variance index (UTFVI). The findings reveal that urban land exhibited a significant expansion trend from 2000 to 2019, reaching 1083.43 km2 in 2019. The forecast indicates that urban land may increase by 8.79% and 10.92% by 2030 and 2040, respectively. Conversely, vegetation and bare land may decrease. The LST is likely to continue to rise, accompanied by a significant expansion of the high temperature range and a contraction of the low temperature range. By 2030 and 2040, the area with LST<20 °C is likely to decrease by 2.17% and 3.19%, while the area with LST>30 °C is likely to expand by 5.68% and 8.08%, respectively. The UTFVI area of urban land may decrease at none and middle levels but may notably expand at stronger and strongest levels. The areas with UTFVI at none, weak, and middle levels show a declining trend, while the increase in UTFVI at the strong level may exceed 46.29% and the strongest level of UTFVI may continue to expand. This study offers new insights into urban sustainable development and thermal environment governance. • Exploring the changes in LULC, LST, and UTFVI in Nanjing over the past 20 years. • Simulating the LULC, LST, and UTFVI in 2030 and 2040 using ANN-CA and ANN algorithms. • Evaluating LULC and thermal environment changes using machine learning algorithms. [ABSTRACT FROM AUTHOR]

Recently, the high demand for marble stones has progressed in the construction industry, ultimately resulting in waste marble production. Thus, environmental degradation is unavoidable because of waste generated from quarry drilling, cutting, and blasting methods. Marble waste is produced in an enormous amount in the form of odd blocks and unwanted rock fragments. Absence of a systematic way to dispose of these marble waste massive mounds results in environmental pollution and landfills. To reduce this risk, an effort has been made for the incorporation of waste marble powder into concrete for sustainable construction. Different proportions of marble powder are considered as a partial substitute in concrete. A total of 40 mixes are prepared. The effectiveness of marble in concrete is assessed by comparing the compressive strength with the plain mix. Supervised machine learning algorithms, bagging (Bg), random forest (RF), AdaBoost (AdB), and decision tree (DT) are used in this study to forecast the compressive strength of waste marble powder concrete. The models' performance is evaluated using correlation coefficient (R2), root mean square error, and mean absolute error and mean square error. The achieved performance is then validated by using the k-fold cross-validation technique. The RF model, having an R2 value of 0.97, has more accurate prediction results than Bg, AdB, and DT models. The higher R2 values and lesser error (RMSE, MAE, and MSE) values are the indicators for better performance of RF model among all individual and ensemble models. The implementation of machine learning techniques for predicting the mechanical properties of concrete would be a practical addition to the civil engineering domain by saving effort, resources, and time. [ABSTRACT FROM AUTHOR]

BACKGROUND Evaluation of the quality properties of papaya becomes essential due to the acceleration of the fruit shelf‐life senescence and the deterioration factor of the expected postharvest operations. In this study, the colour features in RGB, normalised RGB, HSV and L*a*b* channels were extracted and correlated with mechanical properties, moisture content (MC), total soluble solids (TSS) and pH for the prediction of quality properties at five ripening stages of papaya (R1–R5). RESULTS: The mean values of colour features in RGB Rm,Gm,Bm, normalised RGB Rnm,Gnm,Bnm HSV Hm,Sm,Vm, and L*a*b* Lm,am,bm were the best estimator for predicting TSS with R2 ≥ 0.90. All colour channels also showed satisfactory accuracies of R2 ≥ 0.80 in predicting the bioyield force, apparent modulus and mean force. The highest average classification accuracy was obtained using linear discriminant analysis (LDA) with an average accuracy of more than 82%. The study showed that LDA, linear support vector machine, quadratic discriminant analysis and quadratic support vector machine obtained the correct classification of up to 100% for R5, whereas R1, R2, R3 and R4 gave classification accuracies in the range 83.75–91.85%, 85.6–90.25%, 85.75–90.85% and 77.35–87.15%, respectively. This indicates that R5 colour information was obviously different from R1–R4. The mean values of the HSV channel indicated the best performance to predict the ripening stages of papaya, compared to RGB, normalised RGB and L*a*b* channels, with an average classification accuracy of more than 80%. CONCLUSION: The study has shown the versatility of a machine vision system in predicting the quality changes in papaya. The results showed that the machine vision system can be used to predict the ripening stages as well as classifying the fruits into different ripening stages of papayas. © 2021 Society of Chemical Industry. [ABSTRACT FROM AUTHOR]

The Hadamard Extension $\mathbb H({\mathrm {m}})$ of an $n \times k$ matrix m is the collection of all Hadamard products of subsets of its rows. This construction is essential for source identification (parameter estimation) of a mixture of $k$ product distributions over $n$ binary random variables. A necessary requirement for such identification is that $\mathbb H({\mathrm {m}})$ have full column rank; conversely, identification is possible if apart from each row there exist two disjoint sets of rows of m, each of whose extension has full column rank. It is necessary therefore to understand when $\mathbb H({\mathrm {m}})$ has full column rank; we provide two results in this direction. The first is that if $\mathbb H({\mathrm {m}})$ has full column rank then there exists a set of at most $k-1$ rows of m, whose extension already has full column rank. The second is a Hall-type condition on the values in the rows of m, that suffices to ensure full column rank of $\mathbb H({\mathrm {m}})$. [ABSTRACT FROM AUTHOR]

Objective: Diabetes is a chronic fatal disease that has affected millions of people all over the globe. Type 2 Diabetes Mellitus (T2DM) accounts for 90% of the affected population among all types of diabetes. Millions of T2DM patients remain undiagnosed due to lack of awareness and under resourced healthcare system. So, there is a dire need for a diagnostic and prognostic tool that shall help the healthcare providers, clinicians and practitioners with early prediction and hence can recommend the lifestyle changes required to stop the progression of diabetes. The main objective of this research is to develop a framework based on machine learning techniques using only lifestyle indicators for prediction of T2DM disease. Moreover, prediction model can be used without visiting clinical labs and hospital readmissions. Method: A proposed framework is presented and implemented based on machine learning paradigms using lifestyle indicators for better prediction of T2DM disease. The current research has involved different experts like Diabetologists, Endocrinologists, Dieticians, Nutritionists, etc. for selecting the contributing 1552 instances and 11 attributes lifestyle biological features to promote health and manage complications towards T2DM disease. The dataset has been collected through survey and google forms from different geographical regions. Results: Seven machine learning classifiers were employed namely K-Nearest Neighbour (KNN), Linear Regression (LR), Support Vector Machine (SVM), Naive Bayes (NB), Decision Tree (DT), Random Forest (RF) and Gradient Boosting (GB). Gradient Boosting classifier outperformed best with an accuracy rate of 97.24% for training and 96.90% for testing separately followed by RF, DT, NB, SVM, LR, and KNN as 95.36%, 92.52%, 90.72%, 90.20%, 90.20% and 77.06% respectively. However, in terms of precision, RF achieved high performance (0.980%) and KNN performed the lowest (0.793%). As far as recall is being concerned, GB achieved the highest rate of 0.975% and KNN showed the worst rate of 0.774%. Also, GB is top performed in terms of f1-score. According to the ROCs, GB and NB had a better area under the curve compared to the others. Conclusion: The research developed a realistic health management system for T2DM disease based on machine learning techniques using only lifestyle data for prediction of T2DM. To extend the current study, these models shall be used for different, large and real-time datasets which share the commonality of data with T2DM disease to establish the efficacy of the proposed system. [ABSTRACT FROM AUTHOR]

As multipath components (MPCs) are experimentally observed to appear in clusters, cluster-based channel models have been focused in the wireless channel study. However, most of the MPC clustering algorithms for multi-in–multiout (MIMO) channels with delay and angle information of MPCs are based on the distance metric that quantifies the similarity of two MPCs and determines the preferred cluster shape, greatly impacting MPC clustering quality. In this article, a general framework of Mahalanobis-distance metric is proposed for MPC clustering in the MIMO channel analysis, without user-specified parameters. Remarkably, the popular multipath component distance (MCD) is proven to be a special case of the proposed distance metric framework. Furthermore, two machine learning algorithms, namely, weak-supervised Mahalanobis metric for clustering and supervised large margin nearest neighbor, are introduced to learn the distance metric. To evaluate the effectiveness, a modified channel model is proposed based on the Third Generation Partnership Project (3GPP) spatial channel model (SCM) to generate clustered MPCs with delay and angular information since the original 3GPP SCM is incapable to evaluate clustering quality. Experiment results show that the proposed distance metric can significantly improve the clustering quality of existing clustering algorithms, while the learning phase requires considerably limited efforts of labeling MPCs. [ABSTRACT FROM AUTHOR]

In recent past, emerging technology, such as AI has revolutionized progress and advancement of traffic management solutions in context of smart cities. This paper describes a novel approach of using 3D colour‐texture based feature for image detection on public datasets of Chinese traffic sign research database (TSRD) and Mapillary image database. The implementation of 3D colour texture feature for traffic sign detection is evaluated using artificial neural network and multiple other machine learning algorithms as classifier for image detection purposes. Both datasets used in our experiments are considered among most diverse traffic signage datasets globally with annotations of almost all classes. Image datasets have been publicly made available to researchers for academic purposes. For classification of traffic sign images on Chinese TSRD and Mapillary datasets, the result outcome from back‐propagation neural network (NN) classifier outperforms all other and produced best results and under multiple ML algorithms, support vector machines (SVM) cubic has best results. For classification of traffic sign images on Mapillary dataset, the result outcome from rational quadratic Gaussian process regression has best results. Also, the comparisons of results with other similar novel works have also been discussed. The proposed approach outperformed the previous experiments as observed through comparative analysis performed on both datasets. The conclusion of the research work highlights that the image detection performance is noticeably improved by using the combined 3D colour‐texture feature based proposed approach. [ABSTRACT FROM AUTHOR]

Basketball is a mathematical game with many abstract data interpretations. An average fan ceases to witness the revolution in sports, which is influenced using data science and analytics unless someone brings it to light. Nowadays, teams look at data and tend to make decisions on scouting the player for the team. The decision making for the coaches can be made easier using machine learning algorithms to identify the star potential of players. The paper provides a novel algorithm by building a machine learning model on all players to predict whether the player is a star or not. Besides, an interactive user interface is developed for coaches to input the player's data and to make an informed decision based on the prediction. [ABSTRACT FROM AUTHOR]

Industrial Internet of Things (IoT) is the most rapidly developing industry in the current IoT industry, and the intrusion detection system (IDS) remains one of the key technologies for industrial IoT security protection. Researchers have considered applying algorithms such as machine learning and deep learning to network IDSs to cope with complex and changing network environments and to automatically extract key features from high-dimensional feature data. However, in the real industrial IoT environment, data imbalance is the main factor that affects the performance of the deep-learning-based IDS. In this article, we study the network intrusion detection model based on data level. Three data-based research schemes are constructed step by step in this article, which are a data augmentation scheme based on the variational autoencoder (VAE), a data-balancing scheme based on the conditional VAE, and a data-balancing scheme based on random undersampling and conditional VAE. The three data-level-based schemes are combined with the deep-learning-based IDS. In this article, we build experiments based on the CSE-CIC-IDS2018 dataset to verify the effectiveness of three data processing schemes. After data enhancement through the third scheme, the Macro-F1-score of the convolutional-neural-network-based IDS model improved by 3.75% and the Macro-F1-score of the gated-recurrent-unit-based IDS model improved by 5.32%. [ABSTRACT FROM AUTHOR]

Simple Summary: This study adopts a dynamic methodology to explore challenges to the practical application of the adjusted binary classification (ABC) approach, which are related to the unmodifiable characteristics of data used in its development, such as intrasexual variation (sexual dimorphism) of variables and methodological factors such as the selected classification algorithm and sample size. The adequacy of a training dataset's size was judged relative to the classification performance in an independent test set. Finding an optimal classifier was also addressed in this study, wherein the results demonstrate that both statistical modeling and machine learning techniques perform almost equally in the univariate models; however, differences are evident in the multivariate model due to the different number of variables included via the feature selection process, as well as the effect of inadequate training sample size relative to the test set. This approach is particularly useful when quick classification/prediction is required for making real-time forensic decisions. The adjusted binary classification (ABC) approach was proposed to assure that the binary classification model reaches a particular accuracy level. The present study evaluated the ABC for osteometric sex classification using multiple machine learning (ML) techniques: linear discriminant analysis (LDA), boosted generalized linear model (GLMB), support vector machine (SVM), and logistic regression (LR). We used 13 femoral measurements of 300 individuals from a modern Turkish population sample and split data into two sets: training (n = 240) and testing (n = 60). Then, the five best-performing measurements were selected for training univariate models, while pools of these variables were used for the multivariable models. ML classifier type did not affect the performance of unadjusted models. The accuracy of univariate models was 82–87%, while that of multivariate models was 89–90%. After applying ABC to the crossvalidation set, the accuracy and the positive and negative predictive values for uni- and multivariate models were ≥95%. Sex could be estimated for 28–75% of individuals using univariate models but with an obvious sexing bias, likely caused by different degrees of sexual dimorphism and between-group overlap. However, using multivariate models, we minimized the bias and properly classified 81–87% of individuals. A similar performance was also noted in the testing sample (except for FEB), with accuracies of 96–100%, and a proportion of classified individuals between 30% and 82% in univariate models, and between 90% and 91% in multivariate models. When considering different training sample sizes, we demonstrated that LR was the most sensitive with limited sample sizes (n < 150), while GLMB was the most stable classifier. [ABSTRACT FROM AUTHOR]

Spectre attacks can be catastrophic and widespread because they exploit common design flaws caused by the speculative capabilities in modern processors to leak sensitive data through side channels. Completely fixing the problem would require a redesign of the architecture for transient execution or the implementation of a new design on re-configurable hardware. However, such fixes cannot be backported to old machines with fixed hardware design. Completely replacing those machines will take a long time. Moreover, existing software patches may cause significant performance overhead. This paper proposes to detect Spectre by monitoring deviations in microarchitectural events using hardware performance counters with promising accuracy above 90 percent under a variety of workload conditions. However, the attacker may attempt to evade detection by slowing down the attack or mimicking benign programs. This paper thus compares different evasion strategies quantitatively and demonstrates that it is possible for the attacker to avoid detection when operating the attacks at a lower speed while maintaining a reasonable attack success rate. Then, we show that, in order to resist evasion, the original detector must be enhanced by randomly switching between a set of detectors using different features and sampling periods so we can keep the detection accuracy above 80 percent. [ABSTRACT FROM AUTHOR]

Reliable, automatic isotope identification is crucial for improving the performance of radioactivity monitoring, especially in security applications. In this article, a robust identification method based on an ensemble employing artificial neural networks (ANN) was developed and compared with other popular machine learning methods. We have encoded the histogram features using bin-ratio vectors that increase the classification accuracy. To make experimentation more objective, our datasets are generated from real isotope spectra of Cs2LiLaBr6(Ce) (CLLBC) detectors using realistic background and gain shift noise profiles, based on the requirements of American National Standards Institute (ANSI) N42.34. In addition to experimenting with classifying individual isotopes we also evaluate the detection of isotope mixtures, where the proposed method also performs competitively. [ABSTRACT FROM AUTHOR]

Soil water content characteristics included field capacity (FC) and permanent wilting point (PWP) are key soil indicators to study permeability, soil water retention capacity, drainage, irrigation, plant water stress and solute movement. It is necessary to use an appropriate model for predicting these characteristics because of the hardness of their measurement. The present research shows a comparison of intelligent models included ANN, ANFIS, SVM and SVM optimised by firefly and particle swarm meta-heuristic algorithms (SVM-FFA and SVM-PSA) to predict soil water content at −33 and −1500 kPa matric potentials in agricultural soils. Results showed that the SVM-PSA and SVM-FFA hybrid models presented the best performance rather than the others, with predictor variables included geometric mean diameter (dg) of soil particles, bulk density, organic carbon, NDVI and NSMI and the values of R2 and NRMSE in validation dataset, 0.94, 0.081, 0.96 and 0.077 for FC and PWP in SVM-PSA, respectively, and 0.96, 0.079, 0.97 and 0.075 for FC and PWP in SVM-FFA model. Therefore, these intelligent machines are preferable for accurate prediction and could be used to predict soil water content of soils with different soil textures in other regions of Iran or even the world. [ABSTRACT FROM AUTHOR]

Groundwater pollution poses a severe threat and issue to the environment and humanity overall. That is why mitigative strategies are urgently needed. Today, studies mapping groundwater risk pollution assessment are being developed. In this study, five new hybrid/ensemble machine learning (ML) models are developed, named DRASTIC-Random Forest (RF), DRASTIC-Support Vector Machine (SVM), DRASTIC-Multilayer Perceptron (MLP), DRASTIC-RF-SVM, and DRASTIC-RF-MLP, for groundwater pollution assessment in the Saiss basin, in Morocco. The performances of these models are evaluated using the Receiver Operating Characteristic curve (ROC curve), precision, and accuracy. Based on the results of the ROC curve method, it is indicated that the use of hybrid/ensemble machine learning (ML) models improves the performance of the individual machine learning (ML) algorithms. In effect, the AUC value of the original DRASTIC is 0.51. Furthermore, both hybrid/ensemble models, DRASTIC-RF-MLP (AUC = 0.953) and DRASTIC-RF-SVM, (AUC = 0.901) achieve the best accuracy among the other models, followed by DRASTIC-RF (AUC = 0.852), DRASTIC-SVM (AUC = 0.802), and DRASTIC-MLP (AUC = 0.763). The results delineate areas vulnerable to pollution, which require urgent actions and strategies to improve the environmental and social qualities for the local population. [ABSTRACT FROM AUTHOR]

A worm is a standalone program, which is self-replicating malware that distributes itself to other computers and networks. An Internet worm can spread across the network and infect millions of computers in truly little time and the damages caused from such attacks are considered extremely high. In addition, these worms also affect the network packet and its performance, where the packets are analyzed by the signature-based intrusion detection system (IDS) and the network performance is analyzed by the NetFlow based IDS. Hence, this article proposes a joint detection of both the signature based and NetFlow based Internet worms using deep learning convolution neural network (DLCNN) with respect to various attacks and it can also prevent the suspicious actions of attackers (cyber-criminals). Additionally, it provides the security for users' data maintenance, countermeasures, and controls the spreading of the internet worms. The effectiveness of proposed DLCNN model is evaluated using both packet capture (PCAP) and KDD-CUP-99 datasets. Finally, various quality metrics are employed to disclose the superiority of proposed DLCNN model as compared existing machine learning, and back propagated neural network models. [ABSTRACT FROM AUTHOR]

When demonstrating the effectiveness of a new algorithm, researchers are traditionally encouraged to compare their algorithm's performance against existing algorithms on well-studied benchmark test suites. In the absence of more nuanced methodologies, algorithm performance is typically summarized on average across the test suite examples. This paper highlights the potential bias of conclusions drawn by analyzing “on average” performance, and the opportunities offered by a recent testing methodology known as instance space analysis. To illustrate, we revisit our 2007 comparative study of algorithms for facial age estimation, and rigorously stress-test to challenge the original conclusions. The case study demonstrates how powerful visualizations offered by instance space analysis enable greater insights into unique strengths and weaknesses, and which algorithm should be used when and why. Inspired by such insights, a new algorithm is proposed, and its unique advantage is demonstrated. The bias often hidden in well-studied datasets, and the ramifications for drawing biased conclusions, are also illustrated in this case study. While focused on facial age estimation, the methodology and lessons learned from the case study are broadly applicable to any study seeking to draw conclusions about algorithm performance based on empirical results. [ABSTRACT FROM AUTHOR]

Real-time forecasting of the financial time-series data is challenging for many machine learning (ML) algorithms. First, many ML models operate offline, where they need a batch of data, which may not be available during training. Besides, due to a fixed architecture of the majority of the offline-based ML models, they suffer to deal with the uncertain nature of financial time-series data. In contrast, online learning mode evolving-structured ML models could be promising for financial time-series forecasting. For real-time deployment of such models, low memory demand is a must. Besides, the model’s explainability plays a crucial role in forecasting financial time-series. Considering all the requirements, a rule-based autonomous neuro-fuzzy learning algorithm called the parsimonious learning machine (PALM) is proposed here to forecast time-varying stock indices. To provide efficient automation of the proposed algorithm by maintaining the model explainability in terms of limited number linguistic IF-THEN rules, two popular multiobjective evolutionary algorithms (MEAs), such as a real-coded genetic algorithm (GA) and a self-adaptive differential evolution (DE) algorithm are utilized here. In addition, fuzzy type-2 variants of PALMs’ are considered here due to better uncertainty handling capacity than their type-1 counterparts. To evaluate the proposed algorithm’s performance, the closing stock price of fifteen (15) different stock market indices are predicted here. From the results, it is observed that the MEA-based PALMs are performing better than the state-of-the-art benchmark online ML models and providing a rule-based explainable model to the end-user. [ABSTRACT FROM AUTHOR]

Early fault detection in power electronic systems (PESs) to maintain reliability is one of the most important issues that has been significantly addressed in recent years. In this article, after reviewing various works of literature based on fault detection in PESs, data mining-based techniques including artificial neural network, machine learning, and deep learning algorithms are introduced. Then, the fault detection routine in PESs is expressed by introducing signal measurement sensors and how to extract the feature from them. Finally, based on studies, the performance of various data mining methods in detecting PESs faults is evaluated. The results of evaluations show that the deep learning-based techniques given the ability of feature extraction from measured signals are significantly more effective than other methods and as an ideal tool for future applications in the power electronics industry are introduced. [ABSTRACT FROM AUTHOR]

Group Method of Data Handling (GMDH) - type neural network algorithm is the heuristic self-organizing algorithm to model the sophisticated systems. In this study, we propose a new algorithm assembling different classifiers based on GMDH algorithm for binary classification. A Monte Carlo simulation study is conducted to compare diverse classifier ensemble based on GMDH (dce-GMDH) algorithm to the other well-known classifiers and to give recommendations for applied researchers on the selection of appropriate classifier under the different conditions. The simulation study illustrates the proposed approach is more successful than the other classifiers in classification in most scenarios generated under the different conditions. Our proposed method is compared to the other classifiers on Cleveland heart disease data. An implementation of the proposed approach is demonstrated on urine data. Moreover, the proposed algorithm is released under R package GMDH2 under the name of "dceGMDH" for implementation. [ABSTRACT FROM AUTHOR]

Wind speed (WS) is an important factor in wind power generation. Because of this, drastic changes in the WS make it challenging to analyze accurately. Therefore, this study proposed a novel framework based on the stacking ensemble machine learning (SEML) method. The application of a novel framework for WS modeling was developed at sixteen stations in Iran. The SEML method consists of two levels. In particular, eleven machine learning (ML) algorithms in six categories neuron based (artificial neural network (ANN), general regression neural network (GRNN), and radial basis function neural network (RBFNN)), kernel based (least squares support vector machine-grid search (LSSVM-GS)), tree based (M5 model tree (M5), gradient boosted regression (GBR), and least squares boost (LSBoost)), curve based (multivariate adaptive regression splines (MARS)), regression based (multiple linear regression (MLR) and multiple nonlinear regression (MNLR)), and hybrid algorithm based (LSSVM-Harris hawks optimization (LSSVM-HHO)) were selected as the base algorithms in level 1 of the SEML method. In addition, LSBoost was used as a meta-algorithm in level 2 of the SEML method. For this purpose, the output of the base algorithms was used as the input for the LSBoost. A comparison of the results showed that using the SEML method in WS modeling greatly affected the performance of the base algorithms. The highest correlation coefficient (R) in the WS modeling at the sixteen stations using the SEML method was 0.89. The SEML method increased the WS modeling accuracy by >43%. [ABSTRACT FROM AUTHOR]

Large amounts of data, rising costs, and a focus on personalised care have all led to a rise in the use of "big data" in healthcare over the past few years. "Big data processing" is a term used in healthcare to describe the development, collection, analysis, and storage of clinical data that is too big or complicated to be figured out with typical data processing methods. Big data sources for healthcare include the Internet of Things (IoT), Electronic Medical Records/Electronic Health Records (EMR/EHR), which include a patient's medical history, diagnosis, treatment plans, allergies, and lab and test results, and genomic sequencing. In this article, a wide range of healthcare data was analysed using machine learning. As well as the fact that it is hard to gather, handle, and analyse a lot of data. In this paper, we'll take a fresh look at how machine learning algorithms and the need to analyse and use huge amounts of data are linked in real life. Deep learning can be used to look at a lot of data in the health field. [ABSTRACT FROM AUTHOR]

Graphs and other structured data have come to the forefront in machine learning over the past few years due to the efficacy of novel representation learning methods boosting the prediction performance in various tasks. Representation learning methods embed the nodes in a low-dimensional real-valued space, enabling the application of traditional machine learning methods on graphs. These representations have been widely premised to be also suited for graph visualization. However, no benchmarks or encompassing studies on this topic exist. We present an empirical study comparing several state-of-the-art representation learning methods with two recent graph layout algorithms, using readability and distance-based measures as well as the link prediction performance. Generally, no method consistently outperformed the others across quality measures. The graph layout methods provided qualitatively superior layouts when compared to representation learning methods. Embedding graphs in a higher dimensional space and applying t-distributed stochastic neighbor embedding for visualization improved the preservation of local neighborhoods, albeit at substantially higher computational cost. [ABSTRACT FROM AUTHOR]

Class imbalance problem that characterized with the skew distribution towards the majority arises as one challenge in recent years. Many oversampling techniques have been proposed to cope with this problem and some of them combine the oversampling procedure with the clustering algorithm which guaranteeing new synthetic samples being generated in clusters. However far-away samples but with the same minority sub-region are generally clustered into different groups owing to the characteristic of clustering algorithm itself. Therefore, the following oversampling procedure is mostly carried in incomplete minority sub-regions that synthetic samples not well cover the integral minority region. And to our best knowledge, none of existing algorithm is designed to directly estimate minority sub-regions for class imbalance problem. Thus, one new grouping algorithm, named Direction Distribution-based Minority Sub-region Estimation (DDMSE), is first proposed. The new algorithm exploits the intuitive observation, that the minority with the same sub-region almost distribute within the same direction when compared to other majority, to estimate minority sub-regions that tactfully ignoring negative impacts brought by the distance factor like in clustering algorithms. Finally, new synthetic samples are generated in those minority sub-regions. And experimental results on real-world datasets show the comparable performance with other state-of-the-art oversampling methods. [ABSTRACT FROM AUTHOR]

As a promising solution to boost the performance of distance-related algorithms (e.g., $K$ -means and KNN), FPGA-based acceleration attracts lots of attention, but also comes with numerous challenges. In this work, we propose, STPAcc, an optimization framework based on structural triangle-inequality (TI)-based pruning (STP) for accelerating distance-related algorithms on CPU-FPGA platforms. STPAcc provides a domain-specific language to unify distance-related algorithms effectively, a structural TI-based pruning strategy to remove unnecessary distance computations, a coarse-grained workload partitioning and mapping strategy to fully exploit the potentials of the CPU-FPGA platform, and fine-grained hardware optimizations to further improve performance on the FPGA. Intensive experiments show that STPAcc designs achieve $31.42\times $ speedup and $99.63\times $ better energy efficiency on average over standard CPU-based implementations. [ABSTRACT FROM AUTHOR]

Spectral unmixing expresses the mixed pixels existing in hyperspectral images as the product of endmembers and their corresponding fractional abundances, which has been widely used in hyperspectral imagery analysis. However, the endmember spectra even for pixels from the same material of an image may include variability due to the influence of lighting conditions and inherent properties of materials within different pixels. Though the in situ spectral library has been used to accommodate such variability by using multiple in situ spectra to represent each kind of material, the performance improvement may be restricted due to the limited number of endmembers for each material. Therefore, in this article, spectral variability is directly extracted from an in situ endmember library and considered to be transferable among different endmembers for the first time. Furthermore, such a spectral variability is further used to augment sparse unmixing by synchronously performing endmember-based reconstruction and spectral variability-augmented reconstruction in the sparse unmixing model. By, respectively, imposing sparse and smoothness regularization over abundances and variability coefficients, a convex optimization-based spectral variability augmented sparse unmixing (SVASU) is finally proposed, and its convergence performance is also analyzed. Experiments conducted over synthetic and real-world datasets demonstrate that the proposed SVASU method not only significantly improves the unmixing performance of conventional spectral library-based unmixing but also outperforms several state-of-the-art sparse unmixing algorithms. [ABSTRACT FROM AUTHOR]

Objective: Public health authorities monitor epidemiological syndromes to provide early alerts of anomalies. A variety of approaches are applied for effective surveillance systems for influenza like illness (ILI). The present study systematically scores the accuracy of algorithms used for automated and prospective infectious-disease-outbreak detection. Another objective is to improve the performance of machine-learning (ML) approaches through statistical learning. Material and Methods: In order to reflect various situations, the volume and the size of the outbreak is chosen different for each simulation. We simulate 20 yearly sets of "daily ILI visit" to emergency department (ED), which includes seasonal outbreaks as well as unusual outbreaks of varying duration and magnitude. We search which biosurveillance algorithms work best across hidden "unusual outbreaks". Results: In terms of timeliness, both settings of kNN (res-raw), RF (resraw), and LR-raw have the best performance. All ML algorithms have sensitivity results greater than 0.90, where SVM-res (0.97), EWMA (0.96), CUSUM-moderate (0.95) are the best algorithms in terms of specificity. ML algorithms all give better performance with an integrated fitted regression model. The methods which have high sensitivity and specificity together is SVM-res (0.90 and 0.97), and LR-res (0.92 and 0.83). Conclusion: The results verified that ML algorithms integrated with statistical methods can be applied to daily ED data and can be used as a real-time surveillance method for prospective monitoring of ILI cases in the emergency setting. This study can contribute to the early detection of hidden unusual outbreaks for epidemiological studies. [ABSTRACT FROM AUTHOR]

Deep learning is recognized to be capable of discovering deep features for representation learning and pattern recognition without requiring elegant feature engineering techniques by taking advantages of human ingenuity and prior knowledge. Thus it has triggered enormous research activities in machine learning and pattern recognition. One of the most important challenges of deep learning is to figure out relations between a feature and the depth of deep neural networks (deep nets for short) to reflect the necessity of depth. Our purpose is to quantify this feature-depth correspondence in feature extraction and generalization. We present the adaptivity of features to depths and vice-verse via showing a depth-parameter trade-off in extracting both single feature and composite features. Based on these results, we prove that implementing the classical empirical risk minimization on deep nets can achieve the optimal generalization performance for numerous learning tasks. Our theoretical results are verified by a series of numerical experiments including toy simulations and a real application of earthquake seismic intensity prediction. [ABSTRACT FROM AUTHOR]

Highland Andean ecosystems sustain high levels of floral and faunal biodiversity in areas with diverse topography and provide varied ecosystem services, including the supply of water to cities and downstream agricultural valleys. Google (™) has developed a product specifically designed for mapping purposes (Earth Engine), which enables users to harness the computing power of a cloud-based solution in near-real time for land cover change mapping and monitoring. We explore the feasibility of using this platform for mapping land cover types in topographically complex terrain with highly mixed vegetation types (Nor Yauyos Cochas Landscape Reserve located in the central Andes of Peru) using classification machine learning (ML) algorithms in combination with different sets of remote sensing data. The algorithms were trained using 3601 sampling pixels of (a) normalized spectral bands between the visible and near infrared spectrum of the Landsat 8 OLI sensor for the 2018 period, (b) spectral indices of vegetation, soil, water, snow, burned areas and bare ground and (c) topographic-derived indices (elevation, slope and aspect). Six ML algorithms were tested, including CART, random forest, gradient tree boosting, minimum distance, naïve Bayes and support vector machine. The results reveal that ML algorithms produce accurate classifications when spectral bands are used in conjunction with topographic indices, resulting in better discrimination among classes with similar spectral signatures such as pajonal (tussock grass-dominated cover) and short grasses or rocky groups, and moraines, agricultural and forested areas. The model with the highest explanatory power was obtained from the combination of spectral bands and topographic indices using the random forest algorithm (Kappa = 0.81). Our study presents a first approach of its kind in topographically complex Cordilleran terrain and we show that GEE is particularly useful in large-scale land cover mapping and monitoring in mountainous ecosystems subject to rapid changes and conversions, with replicability and scalability to other areas with similar characteristics. [ABSTRACT FROM AUTHOR]

New computational methods have emerged through science and technology to support the diagnosis of mental health disorders. Predictive models developed from machine learning algorithms can identify disorders such as schizophrenia and support clinical decision making. This research aims to compare the performance of machine learning algorithms: Decision Tree, AdaBoost, Random Forest, Naïve Bayes, Support Vector Machine, and k-Nearest Neighbor in the prediction of hospitalized patients with schizophrenia. The data set used in the study contains a total of 11,884 electronic admission records corresponding to 6933 patients with various mental health disorders; these records belong to the acute units of 11 public hospitals in a region of Spain. Of the total, 5968 records correspond to patients diagnosed with schizophrenia (3002 patients) and 5916 records correspond to patients with other mental health disorders (3931 patients). The results recommend Random Forest with the best accuracy of 72.7%. Furthermore, this algorithm presents 79.6%, 72.8%, 72.7%, and 72.7% for AUC, precision, F1-Score, and recall, respectively. The results obtained suggest that the use of machine learning algorithms can classify hospitalized patients with schizophrenia in this population and help in the hospital management of this type of disorder, to reduce the costs associated with hospitalization. [ABSTRACT FROM AUTHOR]

Background: The number of patients with fragility fracture has been increasing. Although the increasing number of patients with fragility fracture increased the rate of fracture (refracture), the causes of refracture are multifactorial, and its predictors are still not clarified. In this issue, we collected a registry-based longitudinal dataset that contained more than 7000 patients with fragility fractures treated surgically to detect potential predictors for clinical refracture. Methods: Based on the fact that machine learning algorithms are often used for the analysis of a large-scale dataset, we developed automatic prediction models and clarified the relevant features for patients with clinical refracture. Formats of input data containing perioperative clinical information were table data. Clinical refracture was documented as the primary outcome if the diagnosis of fracture was made at postoperative outpatient care. A decision-tree-based model, LightGBM, had moderate accuracy for the prediction in the test and the independent dataset, whereas the other models had poor accuracy or worse. Results: From a clinical perspective, rheumatoid arthritis (RA) and chronic kidney disease (CKD) were noted as the relevant features for patients with clinical refracture, both of which were associated with secondary osteoporosis. Conclusion: The decision-tree-based algorithm showed the precise prediction of clinical refracture, in which RA and CKD were detected as the potential predictors. Understanding these predictors may improve the management of patients with fragility fractures. [ABSTRACT FROM AUTHOR]

Detecting the presence of a disease requires laboratory tests, testing kits, and devices; however, these were not always available on hand. This study proposes a new approach in disease detection using machine learning algorithms by analyzing symptoms experienced by a person without requiring laboratory tests. Six supervised machine learning algorithms such as J48 decision tree, random forest, support vector machine, k-nearest neighbors, naïve Bayes algorithms, and artificial neural networks were applied in the "COVID-19 Symptoms and Presence Dataset" from Kaggle. Through hyperparameter optimization and 10-fold cross validation, we attained the highest possible performance of each algorithm. A comparative analysis was performed according to accuracy, sensitivity, specificity, and area under the ROC curve. Results show that random forest, support vector machine, k-nearest neighbors, and artificial neural networks outweighed other algorithms by attaining 98.84% accuracy, 100% sensitivity, 98.79% specificity, and 98.84% area under the ROC curve. Finally, we developed a web application that will allow users to select symptoms currently being experienced, and use it to predict the presence of COVID-19 through the developed prediction model. Based on this mechanism, the proposed method can effectively predict the presence or absence of COVID-19 in a person immediately without using laboratory tests, kits, and devices in a real-time manner. [ABSTRACT FROM AUTHOR]

Algorithms are becoming ubiquitous. However, companies are increasingly alarmed about their algorithms causing major financial or reputational damage. A new industry is envisaged: auditing and assurance of algorithms with the remit to validate artificial intelligence, machine learning, and associated algorithms. [ABSTRACT FROM AUTHOR]

A growing specter in the rise of machine learning is whether the decisions made by machine learning models are fair. While research is already underway to formalize a machine-learning concept of fairness and to design frameworks for building fair models with sacrifice in accuracy, most are geared toward either supervised or unsupervised learning. Yet two observations inspired us to wonder whether semi-supervised learning might be useful to solve discrimination problems. First, previous study showed that increasing the size of the training set may lead to a better trade-off between fairness and accuracy. Second, the most powerful models today require an enormous of data to train which, in practical terms, is likely possible from a combination of labeled and unlabeled data. Hence, in this paper, we present a framework of fair semi-supervised learning in the pre-processing phase, including pseudo labeling to predict labels for unlabeled data, a re-sampling method to obtain multiple fair datasets and lastly, ensemble learning to improve accuracy and decrease discrimination. A theoretical decomposition analysis of bias, variance and noise highlights the different sources of discrimination and the impact they have on fairness in semi-supervised learning. A set of experiments on real-world and synthetic datasets show that our method is able to use unlabeled data to achieve a better trade-off between accuracy and discrimination. [ABSTRACT FROM AUTHOR]