Selected Machine Learning Algorithms Research Articles

Evaluating the Performance of Supervised Machine Learning Algorithms in Breast Cancer Datasets

Breast cancer is the leading cause of mortality globally. Several attempts have been made to use data mining methodology together with machine learning techniques to develop systems that can detect or prevent breast cancer. In line with the reviewed paper; large datasets for illness analysis have been developed. In this study, the results of selected Machine Learning algorithms are compared: Decision Table, J48, SGD, bagging, and Naïve Bayes Updateable on Wisconsin Breast Cancer Original dataset was conducted using weka tools. Exploratory data analysis, pre-processed with supervised attribute selection and class order, was used to identify potential features to aid the performance of the chosen algorithms in classification. The empirical result showed that Decision Table explores greater likelihood (74% correctly classified instances, True Positive Rate of 0.752, False Positive Rate of 0.478, Precision of 0.77, receiver operating characteristic Area of 0.682) in terms of accuracy and efficiency compared with others. This study's comparison technique is thought to aid breast cancer detection.

Machine learning algorithm selection for real-time energy management of hybrid energy ship

The advantages of hybrid energy ship with photovoltaic in energy conservation and emission reduction are becoming more and more prominent with increasing tension of global fossil energy. However, how to deal with photovoltaic uncertainty in real time and make photovoltaic efficiently connected to ship microgrid has become a key technical problem. Therefore, we propose a real-time energy management strategy for hybrid energy ship based on approximate model predictive control. Firstly, aiming at minimizing the operating cost and deviation from the reference state of charge, an energy management framework based on model predictive control is established. Secondly, the machine learning algorithm is trained to approximate the optimal control action of model predictive control offline, and the performance of different machine learning algorithms is analyzed quantitatively. Finally, taking a ferry equipped with photovoltaic as an example, the appropriate machine learning algorithm and sample number are selected. The results show that the proposed strategy can not only ensure the optimization performance, but also effectively reduce the amount of calculation and realize the real-time operation of energy management in hybrid energy ship.

Machine learning and remote sensing techniques applied to estimate soil indicators – Review

• Machine learning to process remote sensing data to estimate soil quality indicators. • Use of unmanned aerial vehicle (UAV) and satellite platforms to estimate soil indicators. • Selection of machine learning algorithms for processing agro-industrial systems databases. • Model for estimating soil quality in agricultural systems at a local scale. The demand for food based on intensive agriculture has decreased soil quality, posing great challenges such as increasing agricultural productivity and promoting environmental sustainability. Thus, researchers have focused on developing models for estimating soil quality based on artificial intelligence techniques for the processing of multidimensional data from agro-industrial systems, which provide useful information for farmers about soil management and crop conditions. However, a model for the application of these new technologies in medium and low-scale agricultural systems has not been identified. Therefore, a review of recent studies of crop yield prediction based on the estimation of chemical, physical, and biological soil quality indicators (SQI), which incorporate different machine learning (ML) techniques to process data from remote sensing (RS) systems, is presented. The advantages and disadvantages are also analyzed for: SQI estimates at regional and local scale, spectral bands used for analysis of plowed soils (bare soils) of cultivation plots, selection of minimun data set (MDS), use of unmanned aerial vehicle (UAV) and satellite platforms, data pre-processing, and selection of ML algorithms for processing biological systems databases (agro-industrial systems). Finally, we present a model to help estimate soil quality in agricultural systems at a local scale, based on ML to process RS data, in the model the inputs to the ML unit come from four different class data sets (RS, SQI, environmental data and crop management data). Crop management uses the production of the ML unit to adjust agricultural management practices and therefore improve crop yield.

The SMART Framework: Selection of Machine Learning Algorithms With ReplicaTions-A Case Study on the Microvascular Complications of Diabetes.

Over 34 million people in the US have diabetes, a major cause of blindness, renal failure, and amputations. Machine learning (ML) models can predict high-risk patients to help prevent adverse outcomes. Selecting the 'best' prediction model for a given disease, population, and clinical application is challenging due to the hundreds of health-related ML models in the literature and the increasing availability of ML methodologies. To support this decision process, we developed the Selection of Machine-learning Algorithms with ReplicaTions (SMART) Framework that integrates building and selecting ML models with decision theory. We build ML models and estimate performance for multiple plausible future populations with a replicated nested cross-validation technique. We rank ML models by simulating decision-maker priorities, using a range of accuracy measures (e.g., AUC) and robustness metrics from decision theory (e.g., minimax Regret). We present the SMART Framework through a case study on the microvascular complications of diabetes using data from the ACCORD clinical trial. We compare selections made by risk-averse, -neutral, and -seeking decision-makers, finding agreement in 80% of the risk-averse and risk-neutral selections, with the risk-averse selections showing consistency for a given complication. We also found that the models that best predicted outcomes in the validation set were those with low performance variance on the testing set, indicating a risk-averse approach in model selection is ideal when there is a potential for high population feature variability. The SMART Framework is a powerful, interactive tool that incorporates various ML algorithms and stakeholder preferences, generalizable to new data and technological advancements.

IoT Framework for Measurement and Precision Agriculture: Predicting the Crop Using Machine Learning Algorithms

IoT architectures facilitate us to generate data for large and remote agriculture areas and the same can be utilized for Crop predictions using this machine learning algorithm. Recommendations are based on the following N, P, K, pH, Temperature, Humidity, and Rainfall these attributes decide the crop to be recommended. The data set has 2200 instances and 8 attributes. Nearly 22 different crops are recommended for a different combination of 8 attributes. Using the supervised learning method, the optimum model is attained using selected machine learning algorithms in WEKA. The Machine learning algorithm selected for classifying is multilayer perceptron rules-based classifier JRip, and decision table classifier. The main objective of this case study is to end up with a model which predicts the high yield crop and precision agriculture. The proposed system modeling incorporates the trending technology, IoT, and Agriculture needy measurements. The performance assessed by the selected classifiers is 98.2273%, the Weighted average Receiver Operator Characteristics is 1 with the maximum time taken to build the model being 8.05 s.

Study on real-time heat release rate inversion for dynamic reconstruction and visualization of tunnel fire scenarios

This study proposes a framework integrating tunnel fire knowledge with the machine learning approach to provide key fire parameters and reconstruct tunnel fire scenarios. The machine learning approach establishes mapping rules between tunnel ceiling temperatures measured by Fiber Bragg Grating (FBG) system and the real-time heat release rate (HRR) based on the tunnel fire database. The database employed in training and testing algorithms was generated by numerical simulations and full-scale tunnel fire tests, respectively. In addition, the HRR database establishment procedures, hyperparameter selection of machine learning algorithms and test dataset selection are systematically analysed. The real-time HRR inversion performance of four machine learning methods (support vector regression (SVR), random forests (RF), multilayer perceptron (MLP), long short-term memory network (LSTM)) and two tunnel fire knowledge methods (maximum ceiling gas temperature, smoke back-layering length) are evaluated. It is shown that the LSTM method provides the best prediction accuracy on full-scale tunnel fire test datasets, and the RF method is found to be most resilient despite the loss of partial temperature data due to the failure of temperature sensors. The corresponding models have been verified by the full-scale tunnel fire tests, and can provide the scientific decision support for intelligent fire-fighting.

The role of data frequency and method selection in electricity price estimation: Comparative evidence from Turkey in pre-pandemic and pandemic periods

The study examines the role of data frequency and estimation methods in electricity price estimation by applying selected machine learning algorithms and time series econometric models. In this context, Turkey is selected as an emerging country example, seven explanatory variables including COVID-19 pandemic is considered, and daily and weekly data between February 20, 2019 and March 26, 2021 that includes pre-pandemic and pandemic periods are used. The empirical results show that (i) machine learning algorithms perform better than time series econometric models for both pre-pandemic and pandemic periods; (ii) high-frequency data increases the performance of estimation models; (iii) machine learning algorithms perform better with high-frequency (daily) data with regard to low-frequency (weekly) data; (iv) the pandemic causes an adverse effect on the performance of estimation models; (v) energy-related variables are more important than other variables although all are significant; (vi) the share of renewable sources in electricity production is the most important variable on the electricity prices in both periods and data types. Hence, the findings highlight the role of data frequency and method selection in electricity prices estimation. Moreover, policy implications are discussed.

Design and Development of an Intelligent System for Predicting 5-Year Survival in Gastric Cancer.

Background:Gastric cancer remains one of the leading causes of worldwide cancer-specific deaths. Accurately predicting the survival likelihood of gastric cancer patients can inform caregivers to boost patient prognostication and choose the best possible treatment path. This study intends to develop an intelligent system based on machine learning (ML) algorithms for predicting the 5-year survival status in gastric cancer patients.Methods:A data set that includes the records of 974 gastric cancer patients retrospectively was used. First, the most important predictors were recognized using the Boruta feature selection algorithm. Five classifiers, including J48 decision tree (DT), support vector machine (SVM) with radial basic function (RBF) kernel, bootstrap aggregating (Bagging), hist gradient boosting (HGB), and adaptive boosting (AdaBoost), were trained for predicting gastric cancer survival. The performance of the used techniques was evaluated with specificity, sensitivity, likelihood ratio, and total accuracy. Finally, the system was developed according to the best model.Results:The stage, position, and size of tumor were selected as the 3 top predictors for gastric cancer survival. Among the 6 selected ML algorithms, the HGB classifier with the mean accuracy, mean specificity, mean sensitivity, mean area under the curve, and mean F1-score of 88.37%, 86.24%, 89.72%, 88.11%, and 89.91%, respectively, gained the best performance.Conclusions:The ML models can accurately predict the 5-year survival and potentially act as a customized recommender for decision-making in gastric cancer patients. The developed system in our study can improve the quality of treatment, patient safety, and survival rates; it may guide prescribing more personalized medicine.

Backtesting comparison of machine learning algorithms with different random seed

Machine learning is nowadays popular area of science research. As the amount of accessible data is still increasing, therefore the machine learning methods can be used for many applications. One of the more detailed topics, which is commonly analyzed in practice in regard to financial decision supporting, is the support for decision making on stock market. It has been noticed that the main focus is on the developed markets like in Asia, West Europe or USA. As the Polish stock market is also recognized as the developed market, here in opposite to Asia, West Europe or USA markets, the shortage of practical implementation of machine learning algorithms is perceptible. With this paper various learning algorithms have been used to determine its backtesting performance. The main goal of this paper is to examine the effectiveness of selected machine learning algorithms and to find the best one for stock data by comparing several selected algorithms using a backtesting environment on the same data sets and general parameters. For this purpose, experiments were carried out on one random seed and then out on 100 different seeds.

Detecting Malware Families and Subfamilies using Machine Learning Algorithms: An Empirical Study

Machine learning algorithms have proved their effectiveness in detecting malware. This paper conducts an em-pirical study to demonstrate the effectiveness of selected machine learning algorithms in detecting and classifying Android malware using permissions features. The used dataset consists of 9000 different malicious applications from the CIC-Maldroid2020, CIC-Maldroid2017 and CIC-InvesAndMal2019 datasets collected by the Canadian Institute for Cybersecurity. Meta-Multiclass and Random Forest ensemble classifiers are used based on different machine learning classifiers to overcome the imbalance in the data classes. Moreover, a genetic attribute selection technique and SMOTE are used to classify Ransomware sub-families to handle the small size of the dataset and underfitting problem. The results show that optimization and ensemble approaches are successful in treating dataset issues, with 95% accuracy in classifying big malware families and 80% in Ransomware subfamilies.

Machine learning applied to acoustic-based road traffic monitoring

The motivation behind this study lies in adapting acoustic noise monitoring systems for road traffic monitoring for driver's safety. Such a system should recognize a vehicle type and weather-related pavement conditions based on the audio level measurement. The study presents the effectiveness of the selected machine learning algorithms in acoustic-based road traffic monitoring. Bases of the operation of the acoustic road traffic detector are briefly described. Principles of several machine learning algorithms, data acquisition process, and information about the dataset built are explained. The study is conducted using the audio recordings prepared by the authors, registered in several locations and under different meteorological conditions of the road surface. For each recording containing a single-vehicle passage, a vector of 67 parameters extracted from the audio signal is calculated. Fisher Linear Discriminant Analysis and Regression Analysis, the fastest among algorithms employed, return the following values of accuracy: 0.968 and 0.978, precision: 0.919 and 0.853, recall: 0.882 and 0.974, and F1-score: 0.898 and 0.868 for vehicle type classification. In the case of the road pavement conditions, the obtained metrics are as follows: accuracy of 0.933, precision of 0.898, recall of 0.9, and F1-score of 0.884.

Framework of Intelligent System for Machine Learning Algorithm Selection in Social Sciences

The ability to generate data has never been as powerful as today when three quintile bytes of data are generated daily. In the field of machine learning, a large number of algorithms have been developed, which can be used for intelligent data analysis and to solve prediction and descriptive problems in different domains. Developed algorithms have different effects on different problems.If one algorithmworks better on one dataset,the same algorithm may work worse on another data set. The reason is that each dataset has different features in terms of local and global characteristics. It is therefore imperative to know intrinsic algorithms behavior on different types of datasets andchoose the right algorithm for the problem solving. To address this problem, this papergives scientific contribution in meta learning field by proposing framework for identifying the specific characteristics of datasets in two domains of social sciences:education and business and develops meta models based on: ranking algorithms, calculating correlation of ranks, developing a multi-criteria model, two-component index and prediction based on machine learning algorithms. Each of the meta models serve as the basis for the development of intelligent system version. Application of such framework should include a comparative analysis of a large number of machine learning algorithms on a large number of datasetsfromsocial sciences.

The Effects of Normalization and Standardization an Internet of Things Attack Detection

It is a known fact that we live in the computer age and that many devices in the world have access to the internet. So how secure are these devices? Is there any guarantee that user information is not accessed from intruder? After the concept of the Internet of Things came into our lives, many things such as seeing the food in our home refrigerator, connecting to the Internet from the car and, and video chatting from our smart watch entered our lives. The number of malicious software is also increasing with these new connections. Researchers are increasingly emphasizing the importance of network security and intensifying their studies. Data preprocessing is very important when designing a secure system. In this study, the importance of normalization and standardization in data preprocessing is examined to make machine learning approaches more successful for detecting attacks on IoT devices. The study was carried out in Logistic Regression, Decision Tree, and Stochastic Gradient Descent machine learning algorithms using the Bot-IoT dataset. Bot-IoT dataset is a popular dataset that is widely used in security studies on IoT devices. Normalization and standardization processes were applied to Bot-IoT dataset separately, so data preprocessing was performed, then selected machine learning algorithms were trained with these -normalized / standardized- datasets. As a result of the trainings made with machine learning algorithms, the values of Accuracy, Precision, Recall and F1 Score rates were examined. And as a result of the study, it was seen that the standardization increased the accuracy rate up to 99.96% in Logistic Regression.

Mobility pattern recognition based prediction for the subway station related bike-sharing trips

The free-floating bike-sharing (BS) system plays an important role in connection with the public transit system. However, few studies have addressed the impacts of the subway network on the BS system and integrated the features quantitatively into the BS trip prediction framework. Based on the observation of the close relationship between the BS and the urban rail transit, our study focuses on the trip forecasting of the BSs around the subway stations. Firstly, the subway station related sites are investigated based on the BS dataset in Beijing, China. Secondly, multiple categories of features are extracted, including the subway station related site categories by clustering, the BS site mobility patterns by tensor decomposition, as well as other features (e.g., temporal, POI, meteorological, and air quality information). Finally, a three-layer ensemble learning model based method (i.e., the SAP-SF method) under the stacking strategy is proposed with integrations of multiple features and the several selected machine learning algorithms. It is applied to the simultaneous prediction of the hourly return numbers for a large-scale BS system involving a total of 280 sites in Beijing. The output performance is also examined by comparing the results with those obtained from the benchmark models. It is indicated that the features of subway station related site categories and site mobility patterns jointly contribute to the improvement of BS trip prediction. The accuracy can be increased layer by layer and is superior to the single machine learning algorithm. The research finding can provide useful information for system administrators to perform service level checks and to rebalance BSs around subway stations.

A Data-Driven Approach to Physical Fatigue Management Using Wearable Sensors to Classify Four Diagnostic Fatigue States.

Physical exercise contributes to the success of rehabilitation programs and rehabilitation processes assisted through social robots. However, the amount and intensity of exercise needed to obtain positive results are unknown. Several considerations must be kept in mind for its implementation in rehabilitation, as monitoring of patients’ intensity, which is essential to avoid extreme fatigue conditions, may cause physical and physiological complications. The use of machine learning models has been implemented in fatigue management, but is limited in practice due to the lack of understanding of how an individual’s performance deteriorates with fatigue; this can vary based on physical exercise, environment, and the individual’s characteristics. As a first step, this paper lays the foundation for a data analytic approach to managing fatigue in walking tasks. The proposed framework establishes the criteria for a feature and machine learning algorithm selection for fatigue management, classifying four fatigue diagnoses states. Based on the proposed framework and the classifier implemented, the random forest model presented the best performance with an average accuracy of ≥98% and F-score of ≥93%. This model was comprised of ≤16 features. In addition, the prediction performance was analyzed by limiting the sensors used from four IMUs to two or even one IMU with an overall performance of ≥88%.

In silico prediction of fragrance retention grades for monomer flavors using QSPR models

The fragrance retention grades (FRGs) of monomer flavors contribute significantly to the perfumer technology development. In silico prediction of FRGs of monomer flavors are required to reduce costs, time, and manual testing. Quantitative structure-property relationships (QSPR) were established employing a database of monomer flavors, including 1552 odorants and corresponding FRGs. Molecular structure physicochemical information of the odorant molecules was acquired using a molecular calculation software (Dragon 7.0). To obviate the challenge of high dimensionality, we employed five feature extractors, including principal component analysis, lasso, recursive feature elimination, autoencoder, and boruta algorithm. Moreover, three machine learning algorithms were applied and compared to develop QSPR models for the estimation of the FRGs of monomer flavors. The selected machine learning algorithms were random forest, support vector machine, and deep neural network. We developed a weighted scoring formula for calculating the correlation score and association analysis between functional groups and FRGs. The results demonstrated that SH (thiols), ArOR (ethers), and ArCOOR (esters) functional groups have significant impact on the FRGs. In addition, we defined the applicability domains (AD) to limit the scope of application of the test dataset and used external data to validate the model reliability. Finally, we performed a comparative analysis using recursive feature elimination to extract the 80-dimensional molecular descriptors (MDs). It was concluded that the random forest algorithm performed better, with an accuracy of 77.81%, precision of 77.83%, recall of 77.99%, and F1-score of 77.88%. The proposed in silico predictive QSPR model is likely to be considered reliable for evaluating the FRGs of monomer flavors and being promoted to the perfumer industry.

Application progress of machine learning in point-of-care testing

<p indent=0mm>In recent years, point-of-care testing (POCT) has attracted more and more attention because it is cheap, fast and easy-to-operate. However, detection accuracy and reliability remain challenging for POCT. Machine learning methods are powerful for data processing and analysis, which is the potential to improve the accuracy and reliability of POCT greatly. Additionally, it is also possible to bring breakthroughs in remote medicine and data sharing fields for POCT. In this review, we describe the basic principles of machine learning algorithms and explain their advantages in POCT; then introduce the applications of machine learning in POCT, including paper-based assays, microfluidic lab-on-chip technologies, and wearable devices; after that, we also put forward suggestions on the selection of machine learning algorithms based on the data type and detection targets of POCT tasks; finally, we propose several directions of the future development of machine learning algorithms in POCT.

Malicious Encryption Traffic Detection Based on NLP

The development of Internet and network applications has brought the development of encrypted communication technology. But on this basis, malicious traffic also uses encryption to avoid traditional security protection and detection. Traditional security protection and detection methods cannot accurately detect encrypted malicious traffic. In recent years, the rise of artificial intelligence allows us to use machine learning and deep learning methods to detect encrypted malicious traffic without decryption, and the detection results are very accurate. At present, the research on malicious encrypted traffic detection mainly focuses on the characteristics’ analysis of encrypted traffic and the selection of machine learning algorithms. In this paper, a method combining natural language processing and machine learning is proposed; that is, a detection method based on TF-IDF is proposed to build a detection model. In the process of data preprocessing, this method introduces the natural language processing method, namely, the TF-IDF model, to extract data information, obtain the importance of keywords, and then reconstruct the characteristics of data. The detection method based on the TF-IDF model does not need to analyze each field of the data set. Compared with the general machine learning data preprocessing method, that is, data encoding processing, the experimental results show that using natural language processing technology to preprocess data can effectively improve the accuracy of detection. Gradient boosting classifier, random forest classifier, AdaBoost classifier, and the ensemble model based on these three classifiers are, respectively, used in the construction of the later models. At the same time, CNN neural network in deep learning is also used for training, and CNN can effectively extract data information. Under the condition that the input data of the classifier and neural network are consistent, through the comparison and analysis of various methods, the accuracy of the one-dimensional convolutional network based on CNN is slightly higher than that of the classifier based on machine learning.

Prediction of desert locust breeding areas using machine learning methods and SMOS (MIR_SMNRT2) Near Real Time product

Despite satellite imagery is being used to identify suitable areas for desert locust, there is a lack of automatized and operational procedures in Near Real Time (NRT). The aim of this study was to assess the capacity of Soil Moisture Near Real Time Neural Network Level 2 product (MIR_SMNRT2) from the Soil Moisture and Ocean Salinity satellite (SMOS) to predict nymphs of desert locust. We used soil moisture time series (between 2016 and 2019) to build 6 machine learning models (logistic regression model “glm”, eXtreme Gradient Boosting “xgbTree”, Weighted k-Nearest Neighbors “kknn”, Feed-Forward Neural Networks and Multinomial Log-Linear Models “nnet”, support vector machine radial “svmRadial”, and random forest “rf”) over the entire recession area. Model results proved that spatial and/or temporal constraints in data sampling conditioned the predictive capacity of the selected machine learning algorithms. Furthermore, we used a forward selection procedure to evaluate the impact that time series data exert on modelling. Our results suggest that soil moisture data retrieved between 95 and 12 days (before the sighting) provided sufficient information to achieve acceptable predictive performances. This methodology can improve current preventive and control operations, it is site-specific, and could be used to other pests.

Benthic habitat classification using multiscale GEOBIA on orthophoto images of Karimunjawa waters

The scale value is an important part of the segmentation stage which is part of Object-Based Image Analysis (OBIA). Selection of scale value can determine the size of the object which affects the results of classification accuracy. In addition to setting the scale value (multiscale), selection of machine learning algorithm applied to classify shallow water benthic habitat objects can also determine the success of the classification. Combination of setting scale values and classification algorithms are aimed to get optimal results by examining classification accuracies. This study uses orthophoto images processed from Unmanned Aerial Vehicle (UAV) mission intended to capture benthic habitat in Karimunjawa waters. The classification algorithms used are Support Vector Machine (SVM), Bayes, and K-Nearest Neighbors (KNN). The results of the classification of combination are then tested for accuracy based on the sample and Training Test Area (TTA) masks. The result shows that SVM algorithm with scale of 300 produces the best level of accuracy. While the lowest accuracy is achieved by using SVM algorithm with scale of 100. The result shows that the optimal scale settings in segmenting objects sequentially are 300, 200, and 100