Supervised Machine Learning Methods Research Articles

Examining Supervised Machine Learning Methods for Integer Link Weight Prediction Using Node Metadata.

With the goal of understanding if the information contained in node metadata can help in the task of link weight prediction, we investigate herein whether incorporating it as a similarity feature (referred to as metadata similarity) between end nodes of a link improves the prediction accuracy of common supervised machine learning methods. In contrast with previous works, instead of normalizing the link weights, we treat them as count variables representing the number of interactions between end nodes, as this is a natural representation for many datasets in the literature. In this preliminary study, we find no significant evidence that metadata similarity improved the prediction accuracy of the four empirical datasets studied. To further explore the role of node metadata in weight prediction, we synthesized weights to analyze the extreme case where the weights depend solely on the metadata of the end nodes, while encoding different relationships between them using logical operators in the generation process. Under these conditions, the random forest method performed significantly better than other methods in 99.07% of cases, though the prediction accuracy was significantly degraded for the methods analyzed in comparison to the experiments with the original weights.

Investigating the Ultrasonic Pulse Velocity of Concrete Containing Waste Marble Dust and Its Estimation Using Artificial Intelligence.

Researchers and engineers are presently focusing on efficient waste material utilization in the construction sector to reduce waste. Waste marble dust has been added to concrete to minimize pollution and landfills problems. Therefore, marble dust was utilized in concrete, and its prediction was made via an artificial intelligence approach to give an easier way to scholars for sustainable construction. Various blends of concrete having 40 mixes were made as partial substitutes for waste marble dust. The ultrasonic pulse velocity of waste marble dust concrete (WMDC) was compared to a control mix without marble dust. Additionally, this research used standalone (multiple-layer perceptron neural network) and supervised machine learning methods (Bagging, AdaBoost, and Random Forest) to predict the ultrasonic pulse velocity of waste marble dust concrete. The models’ performances were assessed using R2, RMSE, and MAE. Then, the models’ performances were validated using k-fold cross-validation. Furthermore, the effect of raw ingredients and their interactions using SHAP analysis was evaluated. The Random Forest model, with an R2 of 0.98, outperforms the MLPNN, Bagging, and AdaBoost models. Compared to all the other models (individual and ensemble), the Random Forest model with greater R2 and lower error (RMSE, MAE) has a superior performance. SHAP analysis revealed that marble dust content has a positive and direct influence on and relationship to the ultrasonic pulse velocity of concrete. Using machine learning to forecast concrete properties saves time, resources, and effort for scholars in the engineering sector.

Cognitive selection of driving automation levels in highly automated vehicles leveraging on Bayesian networking principles

Highly automated vehicles (HAVs) are expected to fundamentally change the on-road transportation field by increasing traffic flow efficiency, reducing road crashes caused by human errors, and increasing comfort and driving productivity. With the help of recent developments in cognitive management techniques and machine learning analysis, intelligent on-board computing services are gaining acceptance. The main goal of the present paper is to introduce and develop a novel on-board cognitive decision-making functionality that dynamically and automatically enables HAVs to operate each time in the best available level of driving automation (LoDA). The proposed functionality utilizes several attributes associated with the driving environment and the driver's personality characteristics and personal preferences. The cognitive nature of the proposed functionality is based on previous knowledge, turned into experience, by implementing the Naive Bayes classifier supervised machine learning method. The effectiveness of the proposed cognitive functionality, in terms of accuracy and speed of convergence, in proactively identifying the optimal LoDA, is illustrated by modelling and analysing three scenarios with regards to drivers with different profile data and to driving scenes with different environment characteristics. Therefore, it can operate as an in-car intelligent personal assistant for the drivers.

Human Resource Information System with Machine Learning Integration

The human resource information system with machine learning integration was developed to aid in the management of employees’ records, profiling, turnover, data analytics, and the generation of electronic personal data sheets used by the government service records. It was developed with the feature of predicting employee turnover using a supervised machine learning method. The system can also generate the following reports, namely, the government service record, years of service and loyalty awards, and available leave credits of the employees. To determine the quality of the developed system the researcher used the ISO 25010 Software Quality Model as a basis when evaluating the properties of a software product. The integration of machine learning in the human resource information system proves to be a very useful tool if integrated into a human resource information system to predict trends in the different aspects of human resource management. Based on the thorough evaluation of the experts and respondents, it was found that the human resource information system is highly usable, secured, efficient, and provides a fast and easy way to manage employees' records and predict employees over using a supervised machine learning that uses the linear regression method.

A Probabilistic Learning Approach Applied to the Optimization of Wake Steering in Wind Farms

Abstract The wake steering control in wind farms has gained significant attention in the last years. This control strategy has shown promise to reduce energy losses due to wake effects and increase the energy production in a wind farm. However, wind conditions are variable in wind farms, and the measurements are uncertain what should be considered in the design of wake steering control strategies. This paper proposes using the probabilistic learning on manifold (PLoM), which can be viewed as a supervised machine learning method, to enable the wake steering optimization under uncertainty. The expected power generation is estimated considering uncertainties in wind speed and direction with good accuracy and reduced computational cost for two wind farm layouts, which expand the application of machine learning models in wake steering. Furthermore, the analysis shows the potential gain with the application of wake steering control.

Revisiting the Biological Behavior of Salmonella enterica in Hydric Resources: A Meta-Analysis Study Addressing the Critical Role of Environmental Water on Food Safety and Public Health.

The increasing number of studies reporting the presence of Salmonella in environmental water sources suggests that it is beyond incidental findings originated from sparse fecal contamination events. However, there is no consensus on the occurrence of Salmonella as its relative serovar representation across non-recycled water sources. We conducted a meta-analysis of proportions by fitting a random-effects model using the restricted maximum-likelihood estimator to obtain the weighted average proportion and between-study variance associated with the occurrence of Salmonella in water sources. Moreover, meta-regression and non-parametric supervised machine learning method were performed to predict the effect of moderators on the frequency of Salmonella in non-recycled water sources. Three sequential steps (identification of information sources, screening and eligibility) were performed to obtain a preliminary selection from identified abstracts and article titles. Questions related to the frequency of Salmonella in aquatic environments, as well as putative differences in the relative frequencies of the reported Salmonella serovars and the role of potential variable moderators (sample source, country, and sample volume) were formulated according to the population, intervention, comparison, and outcome method (PICO). The results were reported according to the Preferred Reporting Items for Systematic Review and Meta-Analyzes statement (PRISMA). A total of 26 eligible papers reporting 148 different Salmonella serovars were retrieved. According to our model, the Salmonella frequency in non-recycled water sources was 0.19 [CI: 0.14; 0.25]. The source of water was identified as the most import variable affecting the frequency of Salmonella, estimated as 0.31 and 0.17% for surface and groundwater, respectively. There was a higher frequency of Salmonella in countries with lower human development index (HDI). Small volume samples of surface water resulted in lower detectable Salmonella frequencies both in high and low HDI regions. Relative frequencies of the 148 serovars were significantly affected only by HDI and volume. Considering that serovars representation can also be affected by water sample volume, efforts toward the standardization of water samplings for monitoring purposes should be considered. Further approaches such as metagenomics could provide more comprehensive insights about the microbial ecology of fresh water and its importance for the quality and safety of agricultural products.

A sparse negative binomial classifier with covariate adjustment for RNA-seq data

Supervised machine learning methods have been increasingly used in biomedical research and clinical practice. In transcriptomic applications, RNA-seq data have become dominating and have gradually replaced traditional microarray, due to their reduced background noise and increased digital precision. Most existing machine learning methods are, however, designed for continuous intensities of microarray and are not suitable for RNA-seq count data. In this paper we develop a negative binomial model via generalized linear model framework with double regularization for gene and covariate sparsity to accommodate three key elements: adequate modeling of count data with overdispersion, gene selection and adjustment for covariate effect. The proposed sparse negative binomial classifier (snbClass) is evaluated in simulations and two real applications of multidisease postmortem brain tissue RNA-seq data and cervical tumor miRNA-seq data to demonstrate its superior performance in prediction accuracy and feature selection.

Machine learning models for accurate pretreatment prediction of chemotherapy associated LV dysfunction in patients with breast cancer and lymphoma receiving chemotherapy (WF-98213 PREVENT and CCCWFU9912 DETECT IV).

1553 Background: Cancer survivors receiving potentially cardiotoxic chemotherapy are at increased risk for developing left ventricular (LV) dysfunction. We implemented machine learning (ML) models to predict future LV dysfunction in patients with breast cancer or lymphoma scheduled to receive potentially cardiotoxic chemotherapy. Methods: We utilized prospectively collected data from NIH studies R01HL118740 (supported by the Wake Forest NCORP Research Base (UG1CA189824)) and R01CA167821. Data included measurements of LV function and demographic factors before, during, and 24 months after initiating potentially cardiotoxic chemotherapy. The two datasets were used both separately and collectively in the development of multiple ML models including penalized linear regression, support vector machine, and random forest (RF). A data preprocessing step properly handled missing information, data imbalance, and encoding. Hyperparameter tuning was performed using cross validation of training data. The final models were assessed with a 20% hold-out test dataset. Cardiotoxicity was defined as a pre- to 24-month post cancer treatment decline in LV ejection fraction (LVEF) of > 10% or to an absolute value of < 50%. Results: 276 patients were included in ML models (7% men, 93% women; age 52±13 years). The RF model based on the combined dataset had the best performance with a prediction accuracy, sensitivity, and specificity of 0.94, 0.81, and 0.98, respectively. The most important variables assessed pre-treatment as measured by the Gini impurity factor were in descending order, LVEF, global LV circumferential strain, LV end-systolic volume, body mass index, LV stroke volume, LV end-diastolic volume, and LV mass. Conclusions: Prior to cancer treatment, supervised ML methods such as RF models predicted declines in LVEF of > 10% and/or to absolute values below 50% would occur 24 months after initiating chemotherapy for breast cancer or lymphoma. With further improvement and validation using larger datasets, these models may play an important role in cardio-oncology care during and following cancer treatment.

Prediction of Host-Pathogen RNA Interaction From RNA Sequences and Dual RNA-seq Data using Variational Autoencoders and Supervised Machine Learning Methods

Prediction of Host-Pathogen RNA Interaction From RNA Sequences and Dual RNA-seq Data using Variational Autoencoders and Supervised Machine Learning Methods

Predicting Van der Waals Heterostructures by a Combined Machine Learning and Density Functional Theory Approach.

Van der Waals (vdW) heterostructures are constructed by different two-dimensional (2D) monolayers vertically stacked and weakly coupled by van der Waals interactions. VdW heterostructures often possess rich physical and chemical properties that are unique to their constituent monolayers. As many 2D materials have been recently identified, the combinatorial configuration space of vdW-stacked heterostructures grows exceedingly large, making it difficult to explore through traditional experimental or computational approaches in a trial-and-error manner. Here, we present a computational framework that combines first-principles electronic structure calculations, 2D material database, and supervised machine learning methods to construct efficient data-driven models capable of predicting electronic and structural properties of vdW heterostructures from their constituent monolayer properties. We apply this approach to predict the band gap, band edges, interlayer distance, and interlayer binding energy of vdW heterostructures. Our data-driven model will open avenues for efficient screening and discovery of low-dimensional vdW heterostructures and moiré superlattices with desired electronic and optical properties for targeted device applications.

Predicting the hydrogen uptake ability of a wide range of zeolites utilizing supervised machine learning methods

This study predicts the hydrogen uptake ability of 28 zeolites using artificial neural networks. The topology-related features of four artificial neural networks and their hyper-parameters determine utilizing 349 experimental data reported for zeolites with different surface areas at wide ranges of pressure and −196 °C. Ranking analysis over various statistical criteria confirms that the most accurate model is the cascade feed-forward neural networks with twelve hidden nodes and logarithm and tangent sigmoid activation functions. This model predicts the experimental data with the absolute average relative deviation of 7.24%, mean absolute error of 0.041, root mean squared error of 0.058, and regression factor of 0.99429. The leverage method approves that 94% of the databank is reliable. Furthermore, the relevancy analysis shows a strong direct relationship between the hydrogen uptake ability of zeolite (HUAZ) and pressure and surface area. This study is the first to develop a model for predicting the HUAZ.

EEG Signal Processing and Supervised Machine Learning to Early Diagnose Alzheimer’s Disease

Electroencephalography (EEG) signal analysis is a fast, inexpensive, and accessible technique to detect the early stages of dementia, such as Mild Cognitive Impairment (MCI) and Alzheimer’s disease (AD). In the last years, EEG signal analysis has become an important topic of research to extract suitable biomarkers to determine the subject’s cognitive impairment. In this work, we propose a novel simple and efficient method able to extract features with a finite response filter (FIR) in the double time domain in order to discriminate among patients affected by AD, MCI, and healthy controls (HC). Notably, we compute the power intensity for each high- and low-frequency band, using their absolute differences to distinguish among the three classes of subjects by means of different supervised machine learning methods. We use EEG recordings from a cohort of 105 subjects (48 AD, 37 MCI, and 20 HC) referred for dementia to the IRCCS Centro Neurolesi “Bonino-Pulejo” of Messina, Italy. The findings show that this method reaches 97%, 95%, and 83% accuracy when considering binary classifications (HC vs. AD, HC vs. MCI, and MCI vs. AD) and an accuracy of 75% when dealing with the three classes (HC vs. AD vs. MCI). These results improve upon those obtained in previous studies and demonstrate the validity of our approach. Finally, the efficiency of the proposed method might allow its future development on embedded devices for low-cost real-time diagnosis.

Evaluation of Sentiment Analysis of Text Using Rule-Based and Automatic Approach

The technique of determining whether a text is good, negative or neutral is known as sentiment analysis (SA).Sentiment Analysis can be identified by many names like Textual Analysis, Opinion Mining. Sentiment Analysis is a branch of Natural Language Processing (NLP) that focuses on the expression of subjective views and feelings about a topic gathered from multiple sources. Sentiment Analysis is a collection of methods for detecting and extracting opinions and uses them for the benefit of business operation. It is a classification algorithm aimed at finding opinions and decision-making point of view. Sentiment Analysis is performed in many ways, Automatic classification approach involves Nave Bayes (NB), Support Vector Machine (SVM), and Linear Regression is examples of supervised machine learning methods (LR). The data is explored using unsupervised machine learning. Recurrent Neural Network (RNN) derivatives are also used for classification. Rule-based approach involves various NLP process for classification.

Distract and Divert: How World Leaders Use Social Media During Contentious Politics

How do leaders communicate during domestic crises? We provide the first global analysis of world leader communication on social media during social unrest. We develop a theory of leaders’ digital communication strategies, building on the diversionary theory of foreign policy, as well as research on the role of democratic institutions in explaining elite responsiveness. To test our theory, we construct a new dataset that characterizes leader communication through social media posts published by any head of state or government on Twitter or Facebook, employing a combination of automated translation and supervised machine learning methods. Our findings show that leaders increase their social media activity and shift the topic from domestic to foreign policy issues during moments of social unrest, which is consistent with a conscious strategy to divert public attention when their position could be at risk. These effects are larger in democracies and in particular in the run-up to elections, which we attribute to incentives created by democratic institutions. Our results demonstrate how social media provide meaningful comparative insight into leaders’ political behavior in the digital age.

Machine learning classification of maladaptive rumination and cognitive distraction in terms of frequency specific complexity

In this study, cognitive and behavioral emotion regulation strategies (ERS) are classified by using machine learning models driven by a new local EEG complexity approach so called Frequency Specific Complexity (FSC) in resting-states (eyes-opened (EO), eyes-closed (EC)). According to international 10–20 electrode placement system, FSC is defined as entropy estimations in Alpha (8-12Hz) and Beta (12.5-30Hz) frequency band intervals of non-overlapped short EEG segments to observe local EEG complexity variations at 62 points on scalp surface. The healthy adults who use both rumination and cognitive distraction frequently are included in the 1st groups, while the others who use these strategies rarely are included in the 2nd group with respect to Cognitive Emotion Regulation Questionnaire (CERQ) scores of them. EEG data and CERQ scores are downloaded from publicly available data-base LEMON. In order to test the reliability of the proposed method, five different supervised machine learning methods in addition to two Extreme Learning Machine models are examined with 5-fold cross-validation for discrimination of the contrasting groups. The highest classification accuracy (CA) of 99.47% is provided by Class-specific Cost Regulation Extreme Learning Machines in EC state. Regarding cortical regions (anterio-frontal, central, temporal, parieto-occipital), the regional FSC estimations did not provide the higher performance, however, corresponding statistical distribution shows the decrease in EEG complexity at mostly anterior cortex in the 1st group characterized by maladaptive rumination. In conclusion, FSC can be proposed to investigate cognitive dysfunctions often caused by the use of rumination.

Hybrid Approach for Wind Turbines Power Curve Modeling Founded on Multi-Agent System and Two Machine Learning Algorithms, K-Means Method and the K-Nearest Neighbors, in the Retrieve Phase of the Dynamic Case Based Reasoning

Wind turbine power curve (WTPC) plays an important role for energy assessment, power forecasting and condition monitoring. The WTPC captures the nonlinear relationship between wind speed and output power. Many modeling approaches have been proposed by researches to improve the WTPC model performance. In this paper, we present a hybrid approach of wind turbines power curve modeling based on Case Based Reasoning approach, multi agent system, the K-Means unsupervised machine learning method, and then the supervised machine learning algorithm, which is the K-Nearest Neighbors KNN method. The both of the Machine Learning algorithms, K-means and KNN, are used in the retrieve step of the Dynamic Case Based Reasoning (DCBR) cycle to facilitate the search of wind turbines with similar characteristics to our target case. These wind turbines are first grouped into homogeneous classes and then sorted on the basis of a feature similarity measure using the K-Nearest Neighbors supervised machine learning method. Finally, a set of WTPC with similar characteristics of the target case are proposed.

Risk Assessment of Polish Joint Stock Companies: Prediction of Penalties or Compensation Payments

Corporate misconduct is a huge and widespread problem in the economy. Many companies make mistakes that result in them having to pay penalties or compensation to other businesses. Some of these cases are so serious that they take a toll on a company’s financial condition. The purpose of this paper was to create and evaluate an algorithm which can predict whether a company will have to pay a penalty and to discover what financial indicators may signal it. The author addresses these questions by applying several supervised machine learning methods. This algorithm may help financial institutions such as banks decide whether to lend money to companies which are not in good financial standing. The research is based on information contained in the financial statements of companies listed on the Warsaw Stock Exchange and NewConnect. Finally, different methods are compared, and methods which are based on gradient boosting are shown to have a higher accuracy than others. The conclusion is that the values of financial ratios can signal which companies are likely to pay a penalty next year.

Abstract 356: Phospholipid Profile Of Cerebral Thrombi In Acute Ischemic Stroke

Background and Purpose: The University of Washington Clot Bank (UWCB) has established a tissue banking protocol optimized for “omic” analysis for cerebral thrombi retrieved during standard-of-care thrombectomy. Thrombi from cerebral arteries of acute ischemic stroke (AIS) patients are a novel resource that allow for the evaluation of lipid biological markers present at the time of ictus. The pathophysiological processes following AIS are complex, and involve ischemia, inflammation, and disturbances of lipid metabolism. The purpose of this study was to characterize the lipid profile from cerebral thrombi in AIS patients. Methods: The UWCB study is a non-probability, convenience sampling of adult subjects (> 18 years of age) who undergo thrombectomy for vascular occlusion. Lipidomics analysis was performed using LC-MS of untargeted lipids on thrombi taken from cerebral arteries of AIS subjects who underwent thrombectomy. Random forest, a supervised machine learning method, was employed to identify lipids associated with cerebral thrombi. Results: Five subjects (age = 53 ± 26.4, 3 males) were included from the UWCB registry. The most common comorbid conditions were atrial fibrillation (40%) and hypertension (40%). The top 10 metabolites identified from the random forest analysis, were of phosphatidylethanolamine molecular species. Phospholipids are involved in structural and functional integrity of cellular membranes and mediate host inflammatory responses. They are a key component of bacterial membranes. Changes in the structure or function of membrane phospholipids of host or relevant human microbiota may lead to alterations in membrane permeability and contribute to pathology of diseases such as neurodegeneration. Conclusion: An untargeted LC-MS lipidomics approach was used to identify lipids in cerebral thrombi at the time of mechanical thrombectomy. The top 10 lipids present in cerebral thrombi were identified, which included significant concentrations of phosphatidylethanolamine species. Because phospholipids play a key regulatory role in complex downstream cellular function, these findings may be of great significance in understanding AIS pathophysiology that has the potential to inform personalized preventive strategies.

Evaluation of Artificial Intelligence Methods to Estimate the Compressive Strength of Geopolymers.

The depletion of natural resources and greenhouse gas emissions related to the manufacture and use of ordinary Portland cement (OPC) pose serious concerns to the environment and human life. The present research focuses on using alternative binders to replace OPC. Geopolymer might be the best option because it requires waste materials enriched in aluminosilicate for its production. The research on geopolymer concrete (GPC) is growing rapidly. However, substantial effort and expenses are required to cast specimens, cures, and tests. Applying novel techniques for the said purpose is the key requirement for rapid and cost-effective research. In this research, supervised machine learning (SML) techniques, including two individual (decision tree (DT) and gene expression programming (GEP)) and two ensembled (bagging regressor (BR) and random forest (RF)) algorithms were employed to estimate the compressive strength (CS) of GPC. The validity and comparison of all the models were made using the coefficient of determination (R2), k-fold, and statistical assessments. It was noticed that the ensembled SML techniques performed better than the individual SML techniques in forecasting the CS of GPC. However, individual SML model results were also in the reasonable range. The R2 value for BR, RF, GEP, and DT models was 0.96, 0.95, 0.93, and 0.88, respectively. The models’ lower error values such as mean absolute error (MAE) and root mean square errors (RMSE) also verified the higher precision of ensemble SML methods. The RF (MAE = 2.585 MPa, RMSE = 3.702 MPa) and BR (MAE = 2.044 MPa, RMSE = 3.180) results are better than the DT (MAE = 4.136 MPa, RMSE = 6.256 MPa) and GEP (MAE = 3.102 MPa, RMSE = 4.049 MPa). The application of SML techniques will benefit the construction sector with fast and cost-effective methods for estimating the properties of materials.

Decision support system for estimating and forecasting state of insurance company

The decision support system was created for estimating and forecasting the state of an insurance company according to its financial and economic indicators. The task of estimating the state of this type of an institution was considered as a problem of a binary classification: whether the company’s activity is efficient or not. During the research, six supervised machine learning methods were implemented: k-nearest neighbors, support vector machine, naive Bayes classifier, random forest, XGBoost and deep neural network. The created system allows the following: to perform correlation analysis of financial and economic indicators, to check the balance of data, to perform training of the selected model and to estimate quality of training, to predict the state of the insurance company according to the selected model. According to the best model, the future state of insurance companies in Ukraine was predicted.