Relevant Features Research Articles

Community-enhanced Link Prediction in Dynamic Networks

The growing popularity of online social networks is quite evident nowadays and provides an opportunity to allow researchers in finding solutions for various practical applications. Link prediction is the technique of understanding network structure and identifying missing and future links in social networks. One of the well-known classes of methods in link prediction is a similarity-based method, which uses local and global topological information of the network to predict missing links. Some methods also exist based on quasi-local features to achieve a trade-off between local and global information on static networks. These quasi-local similarity-based methods are not best suited for considering community information in dynamic networks, failing to balance accuracy and efficiency. Therefore, a community-enhanced framework is presented in this article to predict missing links on dynamic social networks. First, a link prediction framework is presented to predict missing links using parameterized influence regions of nodes and their contribution in community partitions. Then, a unique feature set is generated using local, global, and quasi-local similarity-based as well as community information-based features. This feature set is further optimized using scoring-based feature selection methods to select only the most relevant features. Finally, four machine learning-based classification models are used for link prediction. The experiments are performed on six well-known dynamic networks and three performance metrics, and the results demonstrate that the proposed method outperforms the state-of-the-art methods.

A machine learning model to predict the risk of depression in US adults with obstructive sleep apnea hypopnea syndrome: a cross-sectional study.

Depression is very common and harmful in patients with obstructive sleep apnea hypopnea syndrome (OSAHS). It is necessary to screen OSAHS patients for depression early. However, there are no validated tools to assess the likelihood of depression in patients with OSAHS. This study used data from the National Health and Nutrition Examination Survey (NHANES) database and machine learning (ML) methods to construct a risk prediction model for depression, aiming to predict the probability of depression in the OSAHS population. Relevant features were analyzed and a nomogram was drawn to visually predict and easily estimate the risk of depression according to the best performing model. This is a cross-sectional study. Data from three cycles (2005-2006, 2007-2008, and 2015-2016) were selected from the NHANES database, and 16 influencing factors were screened and included. Three prediction models were established by the logistic regression algorithm, least absolute shrinkage and selection operator (LASSO) algorithm, and random forest algorithm, respectively. The receiver operating characteristic (ROC) area under the curve (AUC), specificity, sensitivity, and decision curve analysis (DCA) were used to assess evaluate and compare the different ML models. The logistic regression model had lower sensitivity than the lasso model, while the specificity and AUC area were higher than the random forest and lasso models. Moreover, when the threshold probability range was 0.19-0.25 and 0.45-0.82, the net benefit of the logistic regression model was the largest. The logistic regression model clarified the factors contributing to depression, including gender, general health condition, body mass index (BMI), smoking, OSAHS severity, age, education level, ratio of family income to poverty (PIR), and asthma. This study developed three machine learning (ML) models (logistic regression model, lasso model, and random forest model) using the NHANES database to predict depression and identify influencing factors among OSAHS patients. Among them, the logistic regression model was superior to the lasso and random forest models in overall prediction performance. By drawing the nomogram and applying it to the sleep testing center or sleep clinic, sleep technicians and medical staff can quickly and easily identify whether OSAHS patients have depression to carry out the necessary referral and psychological treatment.

An efficient and secure wearable sensor based remote healthcare monitoring system using adaptive dilated transformer Bi‐LSTM with gated recurrent unit

AbstractInnovative research works in the healthcare sector keep on advancing every day. As the “Internet of Things (IoT)” keeps on evolving, the application of IoT in the medical field is prominent these days. Utilizing IoT devices, alert messages can be sent directly to medical professionals in case of an emergency. So, monitoring the health condition of an individual using IoT technology has become a popular and beneficial method in today's contemporary medical field. With the help of mobile IoT medical equipment, the technology of smart Healthcare Monitoring System (HMS) is proliferating. By utilizing deep learning and IoT technology, the medical diagnosis system has evolved from direct face‐to‐face visits to the hospital to remote telemedicine method. Most of the data generated by the IoT wearable sensors are highly correlated and may consist of outliers. The extraction of the essential attributes from these data is a complicated task. So, “deep learning and machine learning” techniques are adapted to determine the most relevant and appropriate feature required for efficient diagnosis from the unstructured data produced by the IoT devices and thus help in minimizing the redundancy of unnecessary data. Fusing deep learning methods with healthcare IoT made only the essential details to be available for diagnosis. Therefore, a deep learning‐oriented IoT‐based HMS is executed in this work. With the support of several wearable healthcare devices, the required data are acquired. The encryption of the data acquired from standard sources using Optimal Key‐based Advanced Encryption Standard (OK‐AES) is carried out next to assure the security of the sensitive medical data. The keys for AES encryption are optimally chosen with the aid of the Enhanced Heap‐Based Optimizer Algorithm (EHBOA). The encrypted data is transferred to the “cloud platform” for data storage. Once there is a need for the data, then the encrypted data is initially downloaded from the cloud platform. Then using the same AES scheme, the decryption of the data to attain the original data is carried out. From the retrieved data, the extraction of the crucial attributes is carried out. The extracted features are chosen in an optimized manner and are concatenated with the tuned weights to form the weighted feature matrix. This formulated weighted feature matrix is provided as input to the “Adaptive Dilated Transformer Bidirectional Long Short‐Term Memory (Bi‐LSTM) with Gated Recurrent Unit (GRU) (ADTBi‐LSTM‐GRU) model.” The variables in the ADTBG are optimized using the EHBOA for providing an accurate classification outcome. The classified disease outcome is obtained from the deployed ADTBi‐LSTM‐GRU model. Simulations are done to verify the efficiency and reliability of the implemented deep learning and IoT‐based HMS.

Sulcal variability in anterior lateral prefrontal cortex contributes to variability in reasoning performance among young adults.

Identifying structure-function correspondences is a major goal among biologists, cognitive neuroscientists, and brain mappers. Recent studies have identified relationships between performance on cognitive tasks and the presence or absence of small, shallow indentations, or sulci, of the human brain. Building on the previous finding that the presence of the ventral para-intermediate frontal sulcus (pimfs-v) in the left anterior lateral prefrontal cortex (aLPFC) was related to reasoning task performance in children and adolescents, we tested whether this relationship extended to a different sample, age group, and reasoning task. As predicted, the presence of this aLPFC sulcus was also associated with higher reasoning scores in young adults (ages 22-36). These findings have not only direct developmental, but also evolutionary relevance-as recent work shows that the pimfs-v is exceedingly rare in chimpanzees. Thus, the pimfs-v is a key developmental, cognitive, and evolutionarily relevant feature that should be considered in future studies examining how the complex relationships among multiscale anatomical and functional features of the brain give rise to abstract thought.

InterpretME: A tool for interpretations of machine learning models over knowledge graphs

In recent years, knowledge graphs (KGs) have been considered pyramids of interconnected data enriched with semantics for complex decision-making. The potential of KGs and the demand for interpretability of machine learning (ML) models in diverse domains (e.g., healthcare) have gained more attention. The lack of model transparency negatively impacts the understanding and, in consequence, interpretability of the predictions made by a model. Data-driven models should be empowered with the knowledge required to trace down their decisions and the transformations made to the input data to increase model transparency. In this paper, we propose InterpretME, a tool that using KGs, provides fine-grained representations of trained ML models. An ML model description includes data – (e.g., features’ definition and SHACL validation) and model-based characteristics (e.g., relevant features and interpretations of prediction probabilities and model decisions). InterpretME allows for defining a model’s features over data collected in various formats, e.g., RDF KGs, CSV, and JSON. InterpretME relies on the SHACL schema to validate integrity constraints over the input data. InterpretME traces the steps of data collection, curation, integration, and prediction; it documents the collected metadata in the InterpretME KG. InterpretME is published in GitHub11 https://github.com/SDM-TIB/InterpretME and Zenodo22 https://doi.org/10.5281/zenodo.8112628. The InterpretME framework includes a pipeline for enhancing the interpretability of ML models, the InterpretME KG, and an ontology to describe the main characteristics of trained ML models; a PyPI library of InterpretME is also provided33 https://pypi.org/project/InterpretME/. Additionally, a live code44 https://github.com/SDM-TIB/InterpretME_Demo, and a video55 https://www.youtube.com/watch?v=Bu4lROnY4xg demonstrating InterpretME in several use cases are also available.

Network Security with Virtual Reality based Antivirus Protection and Reduced Detection Delays

Addressing the persistent delay problem in traditional network security antivirus protection systems, this paper introduces an innovative approach utilizing virtual reality (VR) technology. The primary objective is to significantly reduce detection delays and enhance the efficiency of network security measures. An enhanced decision-making algorithm is proposed to identify relevant features associated with network security. These features are then weighted and optimized to improve the overall detection process. An injection list is generated through web crawling techniques to strengthen security measures. A virtual protection block is also developed to serve as a barrier against potential threats. The proposed method claims a detection delay of only 75.33 milliseconds, significantly outperforming two traditional methods that recorded 290.11 milliseconds and 337.30 milliseconds, respectively. This substantial decrease in detection delay emphasizes the effectiveness of automatic detection within the context of VR technology. Practical implementation and empirical evidence further validate the success of this approach. The automatic detection of network security vulnerabilities within the VR technology framework is efficient and exhibits considerable progress. As such, this research offers a promising solution to the delay problem in network security antivirus protection. Embracing VR technology achieves shorter detection delays, ultimately improving the security posture of network systems.

Fault Prediction Using Fuzzy Convolution Neural Network on IOT Environment with Heterogeneous Sensing Data Fusion

Because of the developing worldwide familiarity with natural issues, the expansion of sun-based power plants has turned into an unmistakable element of the energy scene. Nonetheless, keeping up with these sunlights-based offices, especially with regards to recognizing breaking down photovoltaic (PV) cells inside huge scope or far off establishments, presents critical difficulties. The focal goal of our exploration project is to resolve this issue by empowering the convenient identification of flaws in PV cells, consequently possibly saving significant time, exertion, and upkeep costs, especially as for the pivoting gear ordinarily utilized in sun-based power plants. I have developed a non-contact vibration pickup system that makes it possible to collect vibration data from PV cells operating at various speeds and loads without having to physically connect them to machine tools. In addition, I rank and select the most relevant features for accurate fault detection using the Sequential Floating Forward Selection (SFFS) method and Principal Component Analysis (PCA) to reduce the extracted features dimensionality. This thorough methodology offers a promising answer for improve the effectiveness and dependability of sun-oriented power plant upkeep while adding to the more extensive objectives of supportable energy creation and natural safeguarding.

Early prediction of MOOC dropout in self-paced students using deep learning

ABSTRACT To address three issues identified in previous research this study proposes a clustering-based MOOC dropout identification method and an early prediction model based on deep learning. The MOOC learning behavior of self-paced students was analyzed, and two well-known MOOC datasets were used for analysis and validation. The findings are as follows: Firstly, the dropout rate among self-paced students in MOOCs exceeds 90%, with over 50% of students participating in online learning activities for only one day. Furthermore, the starting dates for students in the same course differ significantly. Secondly, leveraging early learning behavior and relevant background features, the proposed early prediction model accurately predicts over 98% of dropout cases and identifies over 50% of engaged students. Through training, the model's convolutional kernels capture meaningful weights for different days and activities. Lastly, background features related to students and courses have a more significant impact on dropout rates. The utilization of resources such as videos and active participation in learning activities, like asking questions, demonstrate a particularly significant influence on dropout rates. Notably, there is no fixed period that consistently affects dropout rates. These method and findings provide effective strategies for decreasing dropout rates and improving student engagement.

Utilization of Facial PPG Signals as a Novel Source for Blood Pressure Estimation

Hypertension, a prevalent global health concern, necessitates accurate and non-invasive blood pressure estimation techniques for effective monitoring and management. This paper proposes a novel machine learning approach utilizing Photo plethysmography (PPG) signals for precise blood pressure estimation. PPG signals, obtained conveniently through wearable devices, offer valuable physiological information related to cardiovascular activity. Leveraging advanced machine learning algorithms, including deep learning architectures and feature extraction methods, our proposed technique aims to establish a robust model for blood pressure estimation using facial image analysis. The methodology involves preprocessing PPG signals, extracting relevant features, and employing sophisticated machine learning models for regression analysis. The evaluation of this novel approach involves comprehensive experimentation with diverse datasets, ensuring its efficacy across various demographic groups and conditions. Results demonstrate promising accuracy and reliability in estimating blood pressure values, suggesting the potential for practical implementation in healthcare settings. The proposed technique showcases a promising avenue for non-invasive and accessible blood pressure monitoring, contributing significantly to personalized healthcare and continuous health monitoring systems.

Digital scanning in orthodontics: a literature review

The landscape of orthodontics is changing, and the transition to digital workflows in clinical practice has begun. As new and updated scanner models are regularly being released, it is paramount to assess the current evidence base and compare this to manufacturer claims. At present, there are limited literature reviews assessing and summarizing the clinically relevant features of dental scanners. This review summarizes key findings with respect to accuracy, patient-related outcomes and clinician experiences with digital dental scanners. CPD/Clinical Relevance: Digital scanners are an increasingly popular device in orthodontics, this article provides an up-to-date summary of the pros and cons of scanners in relation to both the patient and orthodontic clinician.

MECHANICALLY EVOLUTIVE 3D-PRINTED SCAFFOLDS FOR BONE REGENERATION

Developments in the field of additive manufacturing have allowed significant improvements in the design and production of scaffolds with biologically relevant features to treat bone defects. Unfortunately, the workflow to generate personalized scaffolds is source of inaccuracies leading to a poor fit between the implant and patients' bone defects. In addition, scaffolds are often brittle and fragile, uneasing their handling by surgeons, with significant risks of fracture during their insertion in the defect. Consequently, we developed organo-mineral cementitious scaffolds displaying evolutive mechanical properties which are currently being evaluated to treat maxillofacial bone deformities in veterinary clinics. Treatment of dog patients was approved by ethic and welfare committees (CERVO-2022-14-V). To date, 8 puppies with cleft palate/lip deformities received the following treatment. Two weeks prior surgery, CT-scan of patient's skull was performed to allow for surgical planning and scaffold designing. Organo-mineral printable pastes were formulated by mixing an inorganic cement precursor (α-Ca3(PO4)2) to a self-reticulating hydrogel (silanized hyaluronic acid) supplemented with a viscosifier (hydroxymethylpropylcellulose). Scaffolds were produced by robocasting of these pastes. Surgical interventions included the reconstruction of soft tissues, and the insertion of the scaffold soaked with autologous bone marrow. Bone formation was monitored 3 and 6 months after reconstruction, and a biopsy at 6 months was performed for more detailed analyses. Scaffolds displayed great handling properties and were inserted within bone defects without significant issue with a relevant bone edges/scaffold contact. Osteointegration of the scaffolds was observed after 3 months, and regeneration of the defect at 6 months seemed quite promising. Preliminary results have demonstrated a potential of the set-up strategy to treat cleft lip/palate deformities in real, spontaneous clinical setting. Translation of these innovative scaffolds to orthopedics is planned for a near future.

Beef marbling assessment by structured-illumination reflectance imaging with deep learning

Beef marbling is primarily assessed in terms of the abundance and spatial distribution of visible fat flecks within Longissimus Dorsi (LD) muscle. Visual appraisal by trained professionals is the currently prevailing practice in beef marbling assessment, but it is subjective, labor-intensive, and time-consuming. Structured-Illumination Reflectance Imaging (SIRI) offers a potentially useful approach to meat quality assessment as opposed to conventional imaging techniques using uniform illumination. This study presents a new evaluation of a custom-assembled, broadband SIRI system, combined with Deep Learning (DL) approaches, for beef marbling assessment. Beef samples of varying marbling degrees were imaged under sinusoidal illumination at spatial frequencies of 0.05–0.40 cycles/mm. The acquired images were demodulated into Direct Component (DC) and Amplitude Component (AC) images at each spatial frequency. Three DL-based segmentation models, including Unet++, DeepLabv3+, and SegFormer, were built using DC and AC (0.05–0.40 cycles/mm) images for segmenting the LD muscle from the surrounding backfat. Two scenarios of image input, i.e., 1) original beef images without backfat removal and 2) segmented LD muscle images, were examined for feature extraction using a pre-trained ResNeXt-101 model, followed by discriminant modeling to classify beef samples of three marbling categories. SegFormer achieved better LD segmentation than Unet++ and DeepLabv3+ for both DC and AC images, with the best IoU (Intersection over Union) of 96.8% obtained at 0.10 cycles/mm. For marbling classification, the models based on LD segments (scenario 1) consistently produced higher accuracies than those using the original beef images (scenario 2). With the deep features extracted from LD segments, DC images yielded an overall classification accuracy of 83.19%, while the AC images resulted in better accuracies of 85.96%–88.72% (P<0.05), with the greatest improvement of 5.53% (P<0.01) at 0.40 cycles/mm. A further improved accuracy of 90.85% in the marbling classification was achieved by modeling a reduced set of relevant features identified based on MRMR (Minimum Redundancy Maximum Relevance). This study has shown the effectiveness of SIRI for beef marbling assessment.

Classification models combined with Boruta feature selection for heart disease prediction

Cardiovascular disease (CVD), generally called heart illness, is a collective term for various ailments that affect the heart and blood vessels. Heart disease is a primary cause of fatality and morbidity in people worldwide, resulting in 18 million deaths per year. By identifying those who are most vulnerable to heart diseases and ensuring they receive the appropriate care, premature demise can be prevented. Machine learning algorithms are now crucial in the medical field, especially when using medical databases to diagnose diseases. Such efficient algorithms and data processing techniques are applied to predict various diseases and offer much potential for accurate heart disease prognosis. Therefore, this study compares the performance logistic regression, decision tree, and support vector machine (SVM) methods with and without Boruta feature selection. The Cleveland Clinic Heart Disease Dataset acquired from Kaggle, which consists of 14 features and 303 instances, was used for the investigation. It was found that the Boruta feature selection algorithm, which selects six of the most relevant features, improved the results of the algorithms. Among these classification algorithms, logistic regression produced the most efficient result, with an accuracy of 88.52 %.

Novel Darknet traffic data synthesis using Generative Adversarial Networks enhanced with oscillatory Growing Cosine Unit activated convolution layers

The Darknet is an anonymous, encrypted collection of websites, with a passive listening nature - accepting incoming packets, while not supporting outgoing packets. Thus, it can potentially host criminal or malicious activity and software, becoming a cyber security threat. Network Detection Systems are effective in identifying dark net traffic and mitigating its ill effects. However, capturing and extracting data from raw network traffic for training these systems can be time-intensive and costly. Using the CIC-Darknet 2020 dataset, this paper proposes using a Novel Generative Adversarial Networks (GAN) Architecture generating the required training and testing data for these systems. This uses a combination of Growing Cosine Unit (GCU) activated convolution layers and Dense layers for the Generator. Feature selection with statistical correlation methods is used to select the most relevant features. An independently trained Evaluator network is used to evaluate the generated data. The proposed system is compared to other established Tabular data GANs like Conditional Tabular GAN (CTGAN) and Copula GAN with similar parameters and on the same data. The Proposed GAN architecture outperforms CTGAN and CopulaGAN by 20 % and 10 % in ters of accuracy while also taking 90 % and 30 % less time to train respectively. Results from measuring similarity of data using the Inverted Kolmogorov - Smirnov D statistic also show significantly better results for the Proposed GAN Architecture. This shows significant promise in using Generative models to reduce the time and effort costs associated with collecting and formatting data to use in research and for training detection systems.

Indicators for Hospitalization in Acute Pulmonary Embolism: Uncover the Association Between D-dimer Levels, Thrombus Volume and Radiomics

The advent of advanced computed tomography (CT) technology and the field of radiomics has opened up new avenues in diagnostic assessments. Increasingly, there is substantial evidence advocating for the incorporation of quantitative imaging biomarkers in the clinical decision-making process.This study aimed to examine the correlation between D-dimer levels and thrombus size in acute pulmonary embolism (PE) combining dual-energy CT (DECT) and radiomics and to investigate the diagnostic utility of a machine learning classifier based on dual-energy computed tomography (DECT) radiomics for identifying patients with a complicated course, defined as at least hospitalization at IMC. The study was conducted including 136 participants who underwent pulmonary artery CT angiography from January 2015 to March 2022. Based on DECT imaging, 107 radiomic features were extracted for each patient using standardized image processing. After dividing the dataset into training and test sets, stepwise feature reduction based on reproducibility, variable importance and correlation analyses were performed to select the most relevant features; these were used to train and validate the gradient-boosted tree models.Receiver operating characteristics (ROC) analysis was utilized to evaluate the association between volumetric, laboratory data and adverse outcomes. In the central PE group, we observed a significant correlation between thrombus volumetrics and D-dimer levels (p=0.0037), as well as between thrombus volumetrics and hospitalization at the Intermediate Care Unit (IMC) (p=0.0001). In contrast, no statistically significant differences were identified in thrombus sizes between patients who experienced complications and those who had a favorable course (p=0.3162). The trained machine learning classifier achieved an accuracy of 61% and 55% in identifying patients with a complicated course, as indicated by an area under the ROC curve of 0.63 and 0.58. In conclusion, our findings indicate a positive correlation between D-dimer levels and central PE's pulmonary embolic burden. Thrombus volumetrics may serve as an indicator for complications and outcomes in acute PE patients. Thus, thrombus volumetrics, as opposed to D-dimers, could be an additional marker for evaluating embolic disease severity. Moreover, DECT-derived radiomic feature models show promise in identifying patients with a complicated course, such as hospitalization at IMC.

Unboxing Tree ensembles for interpretability: A hierarchical visualization tool and a multivariate optimal re-built tree

The interpretability of models has become a crucial issue in Machine Learning because of algorithmic decisions' growing impact on real-world applications. Tree ensemble methods, such as Random Forests or XgBoost, are powerful learning tools for classification tasks. However, while combining multiple trees may provide higher prediction quality than a single one, it sacrifices the interpretability property resulting in “black-box” models. In light of this, we aim to develop an interpretable representation of a tree-ensemble model that can provide valuable insights into its behavior. First, given a target tree-ensemble model, we develop a hierarchical visualization tool based on a heatmap representation of the forest's feature use, considering the frequency of a feature and the level at which it is selected as an indicator of importance. Next, we propose a mixed-integer linear programming (MILP) formulation for constructing a single optimal multivariate tree that accurately mimics the target model predictions. The goal is to provide an interpretable surrogate model based on oblique hyperplane splits, which uses only the most relevant features according to the defined forest's importance indicators. The MILP model includes a penalty on feature selection based on their frequency in the forest to further induce sparsity of the splits. The natural formulation has been strengthened to improve the computational performance of mixed-integer software. Computational experience is carried out on benchmark datasets from the UCI repository using a state-of-the-art off-the-shelf solver. Results show that the proposed model is effective in yielding a shallow interpretable tree approximating the tree-ensemble decision function.

Explainable Machine Learning for Malware Detection on Android Applications

The presence of malicious software (malware), for example, in Android applications (apps), has harmful or irreparable consequences to the user and/or the device. Despite the protections app stores provide to avoid malware, it keeps growing in sophistication and diffusion. In this paper, we explore the use of machine learning (ML) techniques to detect malware in Android apps. The focus is on the study of different data pre-processing, dimensionality reduction, and classification techniques, assessing the generalization ability of the learned models using public domain datasets and specifically developed apps. We find that the classifiers that achieve better performance for this task are support vector machines (SVM) and random forests (RF). We emphasize the use of feature selection (FS) techniques to reduce the data dimensionality and to identify the most relevant features in Android malware classification, leading to explainability on this task. Our approach can identify the most relevant features to classify an app as malware. Namely, we conclude that permissions play a prominent role in Android malware detection. The proposed approach reduces the data dimensionality while achieving high accuracy in identifying malware in Android apps.

Nondestructive Estimation of Neutral Temperature in Rails: A Comparative Study of Machine Learning Strategies

This paper presents the latest findings of a nondestructive evaluation technique currently under development at the University of Pittsburgh to determine the rail neutral temperature (RNT) in continuous welded rails. The technique is based on the extraction of relevant features from rail vibrations and the use of machine learning (ML) to associate these features to the longitudinal stress of the rail of interest. The features contain the spectral information of the vibrations and are pooled together by frequency domain decomposition for input to ML algorithms. Minimum redundancy–maximum relevance and neighboring component analysis are used to identify relevant features to reduce the size of the input vector. In addition, seven algorithms were considered to identify the most accurate model for neutral temperature with respect to the ground truth RNT measured with a strain-gage rosette. The data used in this study were collected from a 5° curved rail on concrete ties. The vibrations were triggered with a hammer and recorded with a few wireless and wired accelerometers attached on the railhead. The results showed that the Gaussian process regressor performs best, and as few as 20 frequencies can be used to predict the RNT with sufficient accuracy.

Diagnosing diabetes mellitus using machine learning techniques

Diabetes Mellitus (DM) is a frequent condition in which the body's sugar levels are abnormally high for an extended length of time. It is a major cause of death with high mortality rates and the second leading cause of total years lived with disability worldwide. Its seriousness comes from its long-term complications, including nephropathy, retinopathy, and neuropathy leading to kidney failure, poor vision and blindness, and peripheral sensory loss, respectively. Such conditions are life-threatening and affect patients’ quality of life. Therefore, this paper aims to identify the most relevant features in the diagnosis of DM and identify the best classifier that can efficiently diagnose DM based on a set of relevant features. To achieve this, four different feature selection methods have been utilized. Moreover, twelve different classifiers that belong to six learning strategies have been evaluated using two datasets and several evaluation metrics such as Accuracy, Precision, Recall, F1-measure, and ROC area. The obtained results revealed that the correlation attribute evaluation method would be the best choice to handle the task of feature selection and ranking for the considered datasets, especially when considering the Accuracy metric. Furthermore, MultiClassClassifier would be the best classifier to handle Diabetes datasets, especially when considering True Positive, precision, and Recall metrics.

NORMALIZATION AND FEATURE SELECTION USING ENSEMBLE METHODS FOR CROP YIELD PREDICTION

In machine learning study proposes an ensemble-based strategy for both feature selection and data standardization to enhance model performance and interpretability. To maintain consistency across datasets, it employ average filling and weighted K-means clustering. Weighted K-means assigns distinct values to samples based on their distances to cluster centers, offering a more precise representation of the data distribution. Meanwhile, average filling replaces missing values with the average of corresponding features, ensuring a complete dataset for subsequent analysis. For feature selection, adopt an ensemble approach that combines Random Forest (RF) with Logistic Regression (LR) and ElasticNet. RF captures feature importance through tree-based analysis, while LR and ElasticNet provide additional insights into feature relevance and coefficients. This amalgamation aims to provide a comprehensive understanding of feature importance within the dataset. Principal Component Analysis (PCA) is employed to reduce dataset complexity while preserving key properties, facilitating more effective feature selection. By identifying orthogonal components that best explain data variation, PCA enables efficient representation and feature selection. In the final stage, Support Vector Machines (SVM) are utilized for categorization. SVM, a powerful classification method, establishes strong decision boundaries that optimize the gap between classes. Leveraging the selected features, the SVM model effectively categorizes new instances.