Search

Microplastics (MPs) are exudated fragments and fibers from environmental plastic refuse in the lower millimeter to micrometer range with some intentionally created particles and degraded exudates from anthropogenic sources. These particles are often hard to identify due to extensive weathering and material heterogeneity from additives. Leveraging the ability of machine learning (ML) models, which train on features from selected polymer classes, can aid in determining particle identity despite heterogeneity. Herein, a 1,800 spectral training dataset was employed for model training (900 from µ-FTIR and 900 from µ-Raman) using data from common MP polymers (synthetically or naturally derived) and plastics. A threefold battery; synthetic data (SD), mixed synthetic data (MSD), and real-world data from an FTIR Library of Plastic Particles sourced from the Environment (FLOPP-E) and a [Raman] Spectral Library of Plastic Particles aged in the Environment (SLOPP-E); was administered for reliability. Firstly, the SD test determined subspace k-nearest neighbors (SKNN) and wide neural network (WNN) as champions (µ-FTIR and µ-Raman, respectively) with an accuracy of, 99%/100% and 98%/100% (χ2 = 31.99/69, p =.0024/ <.0001). Secondly, the MSD test served as a progenitor to the multi-class prediction from µ-Raman's SKNN showing consistency across 5 replicates (H =.25—6.59, p =.156—.993). And thirdly, the real-world test exhibited a loss in accuracy rate with only the champion, SKNN, retaining ~ 73% and ~ 49% of the correct predictions. Investigation into the probable causes of misprediction via colorant additives led to the discovery of white and blue as the most unpredictable across both databases. Lastly, an investigation of 2 samples revealed confounding colorant additives (SKNN's predictions) as copper phthalocyanine (C2. Blue Fiber) and a derivative from the family of diketo-pyrrolo-pyrroles (Polyester 12. Red Fiber). [ABSTRACT FROM AUTHOR]

This study discusses the impacts of misinformation on social cohesion, trust, and well-being, particularly when targeting specific racial and ethnic groups. It categorizes and reviews various articles to identify the sources and types of misinformation on social media, highlighting common themes and origins. The study briefly acknowledges that generative artificial intelligence and machine learning tools can increase the chance of the generation and spread of harmful misinformation across digital platforms. It also highlights the importance of digital and media literacy education in helping individuals critically evaluate information and navigate online spaces responsibly. Promoting racial and ethnic digital literacy is crucial for protecting against misinformation and fostering informed, representative online discourse. The study calls for a multifaceted approach centered on trust, transparency, and accountability in addressing social media misinformation. Ultimately, it advocates for a culture of critical thinking, factchecking, and ethical behavior to create an online environment that respects diversity, inclusion, and truth, thereby contributing to an informed and empowered society.

The role of pharmacovigilance (PV) in healthcare is to optimize the safety and efficacy of the delivery of pharmaceutical drugs and medical equipment. The increasingly dynamic nature of adverse drug reactions and pharmacovigilance has rendered traditional approaches susceptible to the underreporting of ADRs. Subsequently, the integration of Artificial intelligence in adverse drug reaction reporting is an outstanding technological advancement in pharmacovigilance. Therefore, this critical analysis applied a systematic literature review to comprehend the extensive role of AI in Pharmacovigilance. The research findings acknowledged that AI technologies such as machine learning, deep learning, and natural learning processing (NLP) have automated PV, leading to enhanced signal detection, analysis of unstructured data, risk assessment, and regulatory compliance. [ABSTRACT FROM AUTHOR]

The company's success and growth heavily rely on its workforce's performance, yet the evaluation of employees has been only partially and inconclusively executed so far. The primary goal of this research is to build an open innovation framework for analyzing the performance of the debt collector. We have developed the Reinforcement Learning based Continual Learning (RLC) approach for evaluating the performance by analyzing the metrics such as visit patterns and collection percentage. We have used the private debt collection dataset to assess the debt collector's performance. We formulated hypotheses derived from insights gained during exploratory data analysis and subsequently validated them through statistical testing. Whether there are noticeable distinctions among debt collectors in terms of visitation frequency, collection rates, and collection modes. This proposed open innovation framework for analyzing the debt collector performance provides significant variation in terms of collection rate. The proposed EDQN-CL achieved a 13.56 % higher classification rate than the existing algorithm for categorizing the debt collector performance.

Artificial intelligence (AI) and its subset, machine learning (ML), are growing in acceptance within the space industry. These days, autonomous navigation, spacecraft health monitoring, and operational management of satellite constellations all make extensive use of ML algorithms. However, a lot of research on AI applications in space missions can be categorized into two main categories. One disadvantage of the first category is that it is out of date and ignores a number of noteworthy and recent advances in the field, such as the contributions of deep learning (DL) and bioinspired AI algorithms. The issue with the second category is that it describes specific AI techniques and algorithms in excessive depth, which might be problematic. Owing to these limitations, a concise. [ABSTRACT FROM AUTHOR]

Quantum computing represents a paradigm shift in computational theory, offering unprecedented capabilities for solving complex problems that are intractable for classical computers. The foundational concepts are quantum mechanics, including superposition, entanglement, and quantum gates, these principles form the basis of quantum computation. Quantum bits, or qubits, leverage superposition and entanglement to perform calculations in a fundamentally different manner than classical bits, allowing for exponential parallelism and potentially revolutionary computational power. Quantum machine learning (QML) stands at the nexus of two cutting-edge fields: quantum computing and data science. Combining the principles of quantum mechanics with the computational power of quantum computers, QML aims to harness quantum phenomena to enhance traditional machine learning algorithms. Quantum machine learning algorithms leverage quantum states, such as superposition and entanglement, to process and analyse data in ways that surpass the capabilities of classical methods. Quantum machine learning represents a frontier in both quantum computing and data science, offering novel approaches to data analysis and decision-making with the potential to transform industries and drive scientific discovery. [ABSTRACT FROM AUTHOR]

In the medical field, diagnostic tools that make use of deep neural networks have reached a level of performance never before seen. A proper diagnosis of a patient's condition is crucial in modern medicine since it determines whether or not the patient will receive the care they need. Data from a sinus CT scan is uploaded to a computer and displayed on a high-definition monitor to give the surgeon a clear anatomical orientation before endoscopic sinus surgery. In this study, a unique method is presented for detecting and diagnosing paranasal sinus disorders using machine learning. The researchers behind the current study designed their own approach. To speed up diagnosis, one of the primary goals of our study is to create an algorithm that can accurately evaluate the paranasal sinuses in CT scans. The proposed technology makes it feasible to automatically cut down on the number of CT scan images that require investigators to manually search through them all. In addition, the approach offers an automatic segmentation that may be used to locate the paranasal sinus region and crop it accordingly. As a result, the suggested method dramatically reduces the amount of data that is necessary during the training phase. As a result, this results in an increase in the efficiency of the computer while retaining a high degree of performance accuracy. The suggested method not only successfully identifies sinus irregularities but also automatically executes the necessary segmentation without requiring any manual cropping. This eliminates the need for time-consuming and error-prone human labor. When tested with actual CT scans, the method in question was discovered to have an accuracy of 95.16 percent while retaining a sensitivity of 99.14 percent throughout. [ABSTRACT FROM AUTHOR]

Plant leaf disease identification and classification are the most essential and demanding tasks in the agriculture field. In traditional researches, various automated detection technologies have been developed with the goal of more accurately identifying plant leaf disease. Nevertheless, it faces some problems related to complex mathematical modeling, increased time consumption, processing overhead, and mis-prediction results. Therefore, a novel probabilistic intermittent neural network and artificial jelly fish optimization-based plant leaf disease detection system is proposed in this paper. The proposed work aims to "make a new detection scheme to identify correctly plant leaf disease from the given dataset." Here, the probabilistic intermittent neural network (PINN) classification technique is used to predict label as normal or affected by disease. If it is disease affected, the residual multi-scale Unet segmentation (RMUNet) segmentation technique is applied to segment the disease affected region. Finally, the simulation outcomes confirm the efficiency of the proposed leaf disease identification system under some variables. [ABSTRACT FROM AUTHOR]

Unmanned aerial vehicles (UAVs) are envisioned to play a pivotal role in modern telecommunication and wireless sensor networks, offering unparalleled flexibility and mobility for communication and data collection in diverse environments. This paper presents a comprehensive investigation into the performance of supervised machine learning (ML) models for path loss (PL) prediction in UAV-assisted millimeter-wave (mmWave) radio networks. Leveraging a unique set of interpretable geometrical features, six distinct ML models–linear regression (LR), support vector regressor (SVR), K nearest neighbors (KNN), random forest (RF), extreme gradient boosting (XGBoost), and deep neural network (DNN)–are rigorously evaluated using a massive dataset generated from extensive raytracing (RT) simulations in a typical urban environment. Our results demonstrate that the RF algorithm outperforms other models showcasing superior predictive performance for the test dataset with a root mean square error (RMSE) of 2.38 dB. The proposed ML models demonstrate superior accuracy compared to 3GPP and ITU-R models for mmWave radio networks. This study thoroughly investigates the adaptability of these models to unseen environments and examines the feasibility of training them with sparse datasets to improve accuracy. The reduction in computation time achieved by using ML models instead of extensive RT computations for sparse training datasets is evaluated, and an efficient algorithm for training such models is proposed. Additionally, the sensitivity of ML models to noisy input features is analyzed. We also assess the importance of geometrical features and the impact of sequentially increasing the number of these features on model performance. The results emphasize the significance of the proposed geometrical features and demonstrate the potential of ML models to provide computationally efficient and relatively accurate PL predictions in diverse urban environments.

While academic metrics such as transcripts and GPA are commonly used to evaluate students' knowledge acquisition, there are limited comprehensive metrics to measure their preparedness for the challenges of post-graduation life. This research paper explores the impact of various factors on university students' readiness for change and transition, with a focus on their preparedness for careers. The methodology employed in this study involves designing a survey based on Paul J. Mayer's "The Balance Wheel" to capture students' sentiments on various life aspects, including satisfaction with the educational process and expectations of salary. The collected data from a KBTU student survey (n=47) were processed through machine learning models: Linear Regression, Support Vector Regression (SVR), and Random Forest Regression. Subsequently, an intelligent system was built using these models and fuzzy sets. The system is capable of evaluating graduates' readiness for their future careers and demonstrates a high predictive power. The findings of this research have practical implications for educational institutions. Such an intelligent system can serve as a valuable tool for universities to assess and enhance students' preparedness for post-graduation challenges. By recognizing the factors contributing to students' readiness for change, universities can refine curricula and processes to better prepare students for their career journeys. [ABSTRACT FROM AUTHOR]

Artificial Intelligence (AI) is a field which has spread its wing everywhere. The two subsets of AI – Machine Learning (ML) and Deep Learning (DL) are very much in demand nowadays. Use of AI in healthcare is now seen at a large scale. One of the prominent uses of AI algorithm is in detection of cancer. The number of cancer cases is increasing rapidly and both doctors and researchers are trying to find a solution for the same. Detection of cancer is possible through various tests like MRI, CT scan, etc. It is a challenge for the doctors to interpret the tests correctly as it is very complex and time-consuming. It becomes even more challenging to infer the tests if the cancer is in the early stage. This is where AI can help. Various ML/ DL algorithms can be used to speed up the detection process and thus identify cancerous cells in the initial stage so that the survival rate of the patient increases. Early diagnosis and timely intervention can make a significant difference between life and death for cancer patients. This paper briefly outlines the role of AI including ML and DL in cancer detection and how various researchers are using it for prediction of disease. The paper also outlines the general steps used in AI for detection of cancer. [ABSTRACT FROM AUTHOR]

Due to the exponential increase of internet applications and network users, network traffic classification (NTC) is a crucial study subject. It successfully improves network service identifiability and security concerns of the traffic network and provides a way that improves the Quality of services (QoS). Recently, with the emergence of software-defined networking (SDN) and its ability to get the entire network overview using a centralized controller, machine learning (ML) has been used for NTC. In this paper, an SDN QoS guarantee framework with machine learning traffic classification has been proposed. The framework includes a classification system with two stages, the offline stage, where the classifier was trained and tested, and the online stage, where dealing with the flows and testing the classifier speed is simulated using spark streaming. The result shows that the classifier successfully identifies the specific traffic application with an accuracy of 100% on the "IP-network-traffic-flows-labeled-with-87-apps" dataset and identifies the traffic type with an accuracy of 99.95% on the "ISCX-VPN-NONVPN" dataset. In addition, the classifier speed is proven to be a round 3500 record/sec and a patch duration of 917.3 ms on average with 3210 flows/Trigger. [ABSTRACT FROM AUTHOR]

Background: Endometriosis (EM) is a long-lasting inflammatory disease that is difficult to treat and prevent. Existing research indicates the significance of immune infiltration in the progression of EM. Efferocytosis has an important immunomodulatory function. However, research on the identification and clinical significance of efferocytosis-related genes (EFRGs) in EM is sparse.Methods: The EFRDEGs (differentially expressed efferocytosis-related genes) linked to datasets associated with endometriosis were thoroughly examined utilizing the Gene Expression Omnibus (GEO) and GeneCards databases. The construction of the protein-protein interaction (PPI) and transcription factor (TF) regulatory network of EFRDEGs ensued. Subsequently, machine learning techniques including Univariate logistic regression, LASSO, and SVM classification were applied to filter and pinpoint diagnostic biomarkers. To establish and assess the diagnostic model, ROC analysis, multivariate regression analysis, nomogram, and calibration curve were employed. The CIBERSORT algorithm and single-cell RNA sequencing (scRNA-seq) were employed to explore immune cell infiltration, while the Comparative Toxicogenomics Database (CTD) was utilized for the identification of potential therapeutic drugs for endometriosis. Finally, immunohistochemistry (IHC) and reverse transcription quantitative polymerase chain reaction (RT-qPCR) were utilized to quantify the expression levels of biomarkers in clinical samples of endometriosis.Results: Our findings revealed 13 EFRDEGs associated with EM, and the LASSO and SVM regression model identified six hub genes (ARG2, GAS6, C3, PROS1, CLU, and FGL2). Among these, ARG2, GAS6, and C3 were confirmed as diagnostic biomarkers through multivariate logistic regression analysis. The ROC curve analysis of GSE37837 (AUC = 0.627) and GSE6374 (AUC = 0.635), along with calibration and DCA curve assessments, demonstrated that the nomogram built on these three biomarkers exhibited a commendable predictive capacity for the disease. Notably, the ratio of nine immune cell types exhibited significant differences between eutopic and ectopic endometrial samples, with scRNA-seq highlighting M0 Macrophages, Fibroblasts, and CD8 Tex cells as the cell populations undergoing the most substantial changes in the three biomarkers. Additionally, our study predicted seven potential medications for EM. Finally, the expression levels of the three biomarkers in clinical samples were validated through RT-qPCR and IHC, consistently aligning with the results obtained from the public database.Conclusion: we identified three biomarkers and constructed a diagnostic model for EM in this study, these findings provide valuable insights for subsequent mechanistic research and clinical applications in the field of endometriosis.

Infertility has massively disrupted social and marital life, resulting in stressful emotional well-being. Early diagnosis is the utmost need for faster adaption to respond to these changes, which makes possible via AI tools. Our main objective is to comprehend the role of AI in fertility detection since we have primarily worked to find biomarkers and related risk factors associated with infertility. This paper aims to vividly analyse the role of AI as an effective method in screening, predicting for infertility and related risk factors. Three scientific repositories: PubMed, Web of Science, and Scopus, are used to gather relevant articles via technical terms: (human infertility OR human fertility) AND risk factors AND (machine learning OR artificial intelligence OR intelligent system). In this way, we systematically reviewed 42 articles and performed a meta-analysis. The significant findings and recommendations are discussed. These include the rising importance of data augmentation, feature extraction, explainability, and the need to revisit the meaning of an effective system for fertility analysis. Additionally, the paper outlines various mitigation actions that can be employed to tackle infertility and its related risk factors. These insights contribute to a better understanding of the role of AI in fertility analysis and the potential for improving reproductive health outcomes. [ABSTRACT FROM AUTHOR]

The company's success and growth heavily rely on its workforce's performance, yet the evaluation of employees has been only partially and inconclusively executed so far. The primary goal of this research is to build an open innovation framework for analyzing the performance of the debt collector. We have developed the Reinforcement Learning based Continual Learning (RLC) approach for evaluating the performance by analyzing the metrics such as visit patterns and collection percentage. We have used the private debt collection dataset to assess the debt collector's performance. We formulated hypotheses derived from insights gained during exploratory data analysis and subsequently validated them through statistical testing. Whether there are noticeable distinctions among debt collectors in terms of visitation frequency, collection rates, and collection modes. This proposed open innovation framework for analyzing the debt collector performance provides significant variation in terms of collection rate. The proposed EDQN-CL achieved a 13.56 % higher classification rate than the existing algorithm for categorizing the debt collector performance. [ABSTRACT FROM AUTHOR]

Credit card fraud has adversely impacted market economic order and has broken stakeholders, financial entities, and consumers' trust and interest. Card fraud losses are increasing annually and billions of dollars are being lost. Therefore, this work provides a framework for fraud card detection to be tackled efficiently. Recently, the imbalanced dataset for fraud card transactions due to the number of ordinary transactions being far greater than the amount of fraud. Before solving the fraud problem, we first have to solve the imbalanced data problem which occurred when one class considerably outnumbers the examples of the other class. So, the classification of fraud come to be very tough as the result may get biased towards the majority group. Thus, this paper aims firstly to use hybrid sampling and oversampling preprocessing techniques to solve the imbalanced data problem, and secondly to resolve the fraud. The performance of the proposed framework is estimated based on different metrics accuracy, precision, and recall in comparing existing algorithms such as KNN, LR, LDA, NB, and CART. The obtained results revealing that when the data is highly imbalanced, the model strives to detect fraudulent transactions. Besides, it can predict positive classes improved significantly, reaching an accuracy of 99.9. [ABSTRACT FROM AUTHOR]

Social media is essential in many aspects of our lives. Social media allows us to find news for free. anyone can access it easily at any time. However, social media may also facilitate the rapid spread of misleading news. As a result, there is a probability that low-quality news, including incorrect and fake information, will spread over social media. As well as detecting news credibility on social media becomes essential because fake news can affect society negatively, and the spread of false news has a considerable impact on personal reputation and public trust. In this research, we conducted a model that detects the credibility of Arabic news from social media; particularly Arabic tweets. The content of the tweets revolves around the COVID-19 pandemic. The proposed model applied to detect news credibility using text mining techniques and one of the well-known machine learning classifiers, Decision tree which has the best accuracy equal to 86.6%. [ABSTRACT FROM AUTHOR]

The Internet of Things (IoT) ecosystem has experienced significant growth in data traffic and consequently high dimensionality. Intrusion Detection Systems (IDSs) are essential self-protective tools against various cyber-attacks. However, IoT IDS systems face significant challenges due to functional and physical diversity. These IoT characteristics make exploiting all features and attributes for IDS self-protection difficult and unrealistic. This paper proposes and implements a novel feature selection and extraction approach (i.e., our method) for anomaly-based IDS. The approach begins with using two entropy-based approaches (i.e., information gain (IG) and gain ratio (GR)) to select and extract relevant features in various ratios. Then, mathematical set theory (union and intersection) is used to extract the best features. The model framework is trained and tested on the IoT intrusion dataset 2020 (IoTID20) and NSL-KDD dataset using four machine learning algorithms: Bagging, Multilayer Perception, J48, and IBk. Our approach has resulted in 11 and 28 relevant features (out of 86) using the intersection and union, respectively, on IoTID20 and resulted 15 and 25 relevant features (out of 41) using the intersection and union, respectively, on NSL-KDD. We have further compared our approach with other state-of-the-art studies. The comparison reveals that our model is superior and competent, scoring a very high 99.98% classification accuracy. [ABSTRACT FROM AUTHOR]

The study proposes a novel machine learning (ML) paradigm for cardiovascular disease (CVD) detection in individuals at medium to high cardiovascular risk using data from a Greek cohort of 542 individuals with rheumatoid arthritis, or diabetes mellitus, and/or arterial hypertension, using conventional or office-based, laboratory-based blood biomarkers and carotid/femoral ultrasound image-based phenotypes. Two kinds of data (CVD risk factors and presence of CVD—defined as stroke, or myocardial infarction, or coronary artery syndrome, or peripheral artery disease, or coronary heart disease) as ground truth, were collected at two-time points: (i) at visit 1 and (ii) at visit 2 after 3 years. The CVD risk factors were divided into three clusters (conventional or office-based, laboratory-based blood biomarkers, carotid ultrasound image-based phenotypes) to study their effect on the ML classifiers. Three kinds of ML classifiers (Random Forest, Support Vector Machine, and Linear Discriminant Analysis) were applied in a two-fold cross-validation framework using the data augmented by synthetic minority over-sampling technique (SMOTE) strategy. The performance of the ML classifiers was recorded. In this cohort with overall 46 CVD risk factors (covariates) implemented in an online cardiovascular framework, that requires calculation time less than 1 s per patient, a mean accuracy and area-under-the-curve (AUC) of 98.40% and 0.98 (p < 0.0001) for CVD presence detection at visit 1, and 98.39% and 0.98 (p < 0.0001) at visit 2, respectively. The performance of the cardiovascular framework was significantly better than the classical CVD risk score. The ML paradigm proved to be powerful for CVD prediction in individuals at medium to high cardiovascular risk. [ABSTRACT FROM AUTHOR]

Abstract Background Tumor purity is the percent of cancer cells present in a sample of tumor tissue. The non-cancerous cells (immune cells, fibroblasts, etc.) have an important role in tumor biology. The ability to determine tumor purity is important to understand the roles of cancerous and non-cancerous cells in a tumor. Methods We applied a supervised machine learning method, XGBoost, to data from 33 TCGA tumor types to predict tumor purity using RNA-seq gene expression data. Results Across the 33 tumor types, the median correlation between observed and predicted tumor-purity ranged from 0.75 to 0.87 with small root mean square errors, suggesting that tumor purity can be accurately predicted υσινγ expression data. We further confirmed that expression levels of a ten-gene set (CSF2RB, RHOH, C1S, CCDC69, CCL22, CYTIP, POU2AF1, FGR, CCL21, and IL7R) were predictive of tumor purity regardless of tumor type. We tested whether our set of ten genes could accurately predict tumor purity of a TCGA-independent data set. We showed that expression levels from our set of ten genes were highly correlated (ρ = 0.88) with the actual observed tumor purity. Conclusions Our analyses suggested that the ten-gene set may serve as a biomarker for tumor purity prediction using gene expression data.

Medical image segmentation is a key step to assist diagnosis of several diseases, and accuracy of a segmentation method is important for further treatments of different diseases. Different medical imaging modalities have different challenges such as intensity inhomogeneity, noise, low contrast, and ill-defined boundaries, which make automated segmentation a difficult task. To handle these issues, we propose a new fully automated method for medical image segmentation, which utilizes the advantages of thresholding and an active contour model. In this study, a Harris Hawks optimizer is applied to determine the optimal thresholding value, which is used to obtain the initial contour for segmentation. The obtained contour is further refined by using a spatially varying Gaussian kernel in the active contour model. The proposed method is then validated using a standard skin dataset (ISBI 2016), which consists of variable-sized lesions and different challenging artifacts, and a standard cardiac magnetic resonance dataset (ACDC, MICCAI 2017) with a wide spectrum of normal hearts, congenital heart diseases, and cardiac dysfunction. Experimental results show that the proposed method can effectively segment the region of interest and produce superior segmentation results for skin (overall Dice Score 0.90) and cardiac dataset (overall Dice Score 0.93), as compared to other state-of-the-art algorithms. [ABSTRACT FROM AUTHOR]

The Internet of Things (IoT) ecosystem has experienced significant growth in data traffic and consequently high dimensionality. Intrusion Detection Systems (IDSs) are essential self-protective tools against various cyber-attacks. However, IoT IDS systems face significant challenges due to functional and physical diversity. These IoT characteristics make exploiting all features and attributes for IDS self-protection difficult and unrealistic. This paper proposes and implements a novel feature selection and extraction approach (i.e., our method) for anomaly-based IDS. The approach begins with using two entropy-based approaches (i.e., information gain (IG) and gain ratio (GR)) to select and extract relevant features in various ratios. Then, mathematical set theory (union and intersection) is used to extract the best features. The model framework is trained and tested on the IoT intrusion dataset 2020 (IoTID20) and NSL-KDD dataset using four machine learning algorithms: Bagging, Multilayer Perception, J48, and IBk. Our approach has resulted in 11 and 28 relevant features (out of 86) using the intersection and union, respectively, on IoTID20 and resulted 15 and 25 relevant features (out of 41) using the intersection and union, respectively, on NSL-KDD. We have further compared our approach with other state-of-the-art studies. The comparison reveals that our model is superior and competent, scoring a very high 99.98% classification accuracy.

Reliable credit scoring models played a very important role of retail banks to evaluate credit applications and it has been widely studied. The main objective of this paper is to build a hybrid credit scoring model using feature selection approach. In this study, we constructed a credit scoring model based on parallel GBM (Gradient Boosted Model), filter and wrapper approaches to evaluate the applicant’s credit score from the input features. Feature scoring expression are combined by feature important (Gini index) and Information Value. Backward sequential scheme is used for selecting optimal subset of relevant features while the subset is evaluated by GBM classifier. To reduce the running time, we applied parallel GBM classifier to evaluate the proposed subset of features. The experimental results showed that the proposed method obtained a higher predictive accuracy than a baseline method for some certain datasets. It also showed faster speed and better generalization than traditional feature selection methods widely used in credit scoring.

Background: Tumor purity is the percent of cancer cells present in a sample of tumor tissue. The non-cancerous cells (immune cells, fibroblasts, etc.) have an important role in tumor biology. The ability to determine tumor purity is important to understand the roles of cancerous and non-cancerous cells in a tumor. Methods: We applied a supervised machine learning method, XGBoost, to data from 33 TCGA tumor types to predict tumor purity using RNA-seq gene expression data. Results: Across the 33 tumor types, the median correlation between observed and predicted tumor-purity ranged from 0.75 to 0.87 with small root mean square errors, suggesting that tumor purity can be accurately predicted υσινγ expression data. We further confirmed that expression levels of a ten-gene set (CSF2RB, RHOH, C1S, CCDC69, CCL22, CYTIP, POU2AF1, FGR, CCL21, and IL7R) were predictive of tumor purity regardless of tumor type. We tested whether our set of ten genes could accurately predict tumor purity of a TCGA-independent data set. We showed that expression levels from our set of ten genes were highly correlated (ρ = 0.88) with the actual observed tumor purity. Conclusions: Our analyses suggested that the ten-gene set may serve as a biomarker for tumor purity prediction using gene expression data. [ABSTRACT FROM AUTHOR]

Abstract—: Here we provide a short description of an expert system for predicting the ice-jamming power in the area of the Northern Dvina River and a procedure to verify this system. This expert system is based on hydrological and meteorological data for 1991-2016. The data were processed using a machine learning technique and adjacent mathematics, because there is no mathematical model of the ice-jamming process and time series of observations are too short to apply classical statistics. This expert system was developed in 2012; it was adjusted using data from 1991-2010 seasons obtained at hydrological stations. The current investigation involves additional data on 2011-2016 seasons to repeat learning and estimate system quality. The developed system demonstrates a reliable efficiency: the forecast results coincide with observations for all six added seasons (2011-2016). It should be noted that the additional data do not change forecast accuracy, which remained approximately 85%, like in the previous study. All developed software is cross-platform, written with a C++ language, and is implemented as a command line application. This software can be easily adopted to operate as a part of the Northern Dvina River online monitoring service. [ABSTRACT FROM AUTHOR]

Traces collected by citizens using GNSS (Global Navigation Satellite System) devices during sports activities such as running, hiking or biking are now widely available through different sport-oriented collaborative websites. The traces are collected by citizens for their own purposes and frequently shared with the sports community on the internet. Our research assumption is that crowdsourced GNSS traces may be a valuable source of information to detect updates in authoritative datasets. Despite their availability, the traces present some issues such as poor metadata, attribute incompleteness and heterogeneous positional accuracy. Moreover, certain parts of the traces (GNSS points composing the traces) are results of the displacements made out of the existing paths. In our context (i.e., update authoritative data) these off path GNSS points are considered as noise and should be filtered. Two types of noise are examined in this research: Points representing secondary activities (e.g., having a lunch break) and points representing errors during the acquisition. The first ones we named secondary human behaviour (SHB), whereas we named the second ones outliers. The goal of this paper is to improve the smoothness of traces by detecting and filtering both SHB and outliers. Two methods are proposed. The first one allows for the detection secondary human behaviour by analysing only traces geometry. The second one is a rule-based machine learning method that detects outliers by taking into account the intrinsic characteristics of points composing the traces, as well as the environmental conditions during traces acquisition. The proposed approaches are tested on crowdsourced GNSS traces collected in mountain areas during sports activities.

Traces collected by citizens using GNSS (Global Navigation Satellite System) devices during sports activities such as running, hiking or biking are now widely available through different sport-oriented collaborative websites. The traces are collected by citizens for their own purposes and frequently shared with the sports community on the internet. Our research assumption is that crowdsourced GNSS traces may be a valuable source of information to detect updates in authoritative datasets. Despite their availability, the traces present some issues such as poor metadata, attribute incompleteness and heterogeneous positional accuracy. Moreover, certain parts of the traces (GNSS points composing the traces) are results of the displacements made out of the existing paths. In our context (i.e., update authoritative data) these off path GNSS points are considered as noise and should be filtered. Two types of noise are examined in this research: Points representing secondary activities (e.g., having a lunch break) and points representing errors during the acquisition. The first ones we named secondary human behaviour (SHB), whereas we named the second ones outliers. The goal of this paper is to improve the smoothness of traces by detecting and filtering both SHB and outliers. Two methods are proposed. The first one allows for the detection secondary human behaviour by analysing only traces geometry. The second one is a rule-based machine learning method that detects outliers by taking into account the intrinsic characteristics of points composing the traces, as well as the environmental conditions during traces acquisition. The proposed approaches are tested on crowdsourced GNSS traces collected in mountain areas during sports activities. [ABSTRACT FROM AUTHOR]