Machine Learning-based Classification Research Articles

Cataract Eye Disease Diagnosis Using the Random Forest Method

This study developed a machine learning-based classification model using the Random Forest algorithm to detect cataract risk based on 11 variables: age, gender, family history, lens opacity, visual acuity reduction, light sensitivity, color changes, double vision, intraocular pressure, slit-lamp results, and visual acuity. Feature importance analysis revealed that lens opacity and visual acuity variables contributed most significantly to cataract risk prediction, followed by intraocular pressure and visual acuity reduction. The system was designed using Google Colab for model training and Streamlit as an interactive interface, enabling real-time predictions with intuitive result visualization. After optimization using Grid Search, the model achieved an accuracy of 92.0%, precision of 95.0%, sensitivity of 90.0%, F1 Score of 92.4%, and specificity of 98.0%. This system is expected to serve as an effective supporting tool for medical professionals in the early diagnosis of cataracts.

HViTML: Hybrid vision transformer with machine learning-based classification model for glaucomatous eye

HViTML: Hybrid vision transformer with machine learning-based classification model for glaucomatous eye

Machine Learning Algorithms in Controlled Donation After Circulatory Death Under Normothermic Regional Perfusion: A Graft Survival Prediction Model.

Several scores have been developed to stratify the risk of graft loss in controlled donation after circulatory death (cDCD). However, their performance is unsatisfactory in the Spanish population, where most cDCD livers are recovered using normothermic regional perfusion (NRP). Consequently, we explored the role of different machine learning-based classifiers as predictive models for graft survival. A risk stratification score integrated with the model of end-stage liver disease score in a donor-recipient (D-R) matching system was developed. This retrospective multicenter cohort study used 539 D-R pairs of cDCD livers recovered with NRP, including 20 donor, recipient, and NRP variables. The following machine learning-based classifiers were evaluated: logistic regression, ridge classifier, support vector classifier, multilayer perceptron, and random forest. The endpoints were the 3- and 12-mo graft survival rates. A 3- and 12-mo risk score was developed using the best model obtained. Logistic regression yielded the best performance at 3 mo (area under the receiver operating characteristic curve = 0.82) and 12 mo (area under the receiver operating characteristic curve = 0.83). A D-R matching system was proposed on the basis of the current model of end-stage liver disease score and cDCD-NRP risk score. The satisfactory performance of the proposed score within the study population suggests a significant potential to support liver allocation in cDCD-NRP grafts. External validation is challenging, but this methodology may be explored in other regions.

Machine learning-based classification and prediction of typical Chinese green tea taste profiles

Machine learning-based classification and prediction of typical Chinese green tea taste profiles

Congruent Feature Selection Method to Improve the Efficacy of Machine Learning-Based Classification in Medical Image Processing

Congruent Feature Selection Method to Improve the Efficacy of Machine Learning-Based Classification in Medical Image Processing

Machine learning-based classification of medication adherence among patients with noncommunicable diseases

Machine learning-based classification of medication adherence among patients with noncommunicable diseases

Sentiment Analysis of Product Reviews Using Machine Learning and Pre-Trained LLM

Sentiment analysis via artificial intelligence, i.e., machine learning and large language models (LLMs), is a pivotal tool that classifies sentiments within texts as positive, negative, or neutral. It enables computers to automatically detect and interpret emotions from textual data, covering a spectrum of feelings without direct human intervention. Sentiment analysis is integral to marketing research, helping to gauge consumer emotions and opinions across various sectors. Its applications span analyzing movie reviews, monitoring social media, evaluating product feedback, assessing employee sentiments, and identifying hate speech. This study explores the application of both traditional machine learning and pre-trained LLMs for automated sentiment analysis of customer product reviews. The motivation behind this work lies in the demand for more nuanced understanding of consumer sentiments that can drive data-informed business decisions. In this research, we applied machine learning-based classifiers, i.e., Random Forest, Naive Bayes, and Support Vector Machine, alongside the GPT-4 model to benchmark their effectiveness for sentiment analysis. Traditional models show better results and efficiency in processing short, concise text, with SVM in classifying sentiment of short length comments. However, GPT-4 showed better results with more detailed texts, capturing subtle sentiments with higher precision, recall, and F1 scores to uniquely identify mixed sentiments not found in the simpler models. Conclusively, this study shows that LLMs outperform traditional models in context-rich sentiment analysis by not only providing accurate sentiment classification but also insightful explanations. These results enable LLMs to provide a superior tool for customer-centric businesses, which helps actionable insights to be derived from any textual data.

Machine Learning Methods in IoT Based Embedded Systems for Classifying Physical Faults in Water Distribution Networks

Water is the most important factor for the survival of living things on Earth. Although 70% of the Earth is water, the amount of drinkable water is approximately 0.3%. Therefore, creating a sustainable water policy and carrying out studies are very important for our world and our future. Most of the potable water resources are physical losses. In the evaluations made based on metropolitan municipalities, it was seen that the water loss rate was approximately 50%. The study aims to find water pipe faults using IoT (Internet of Things) based machine learning classifiers to prevent physical losses in water distribution networks. Within the scope of this study, an experimental environment was created and an IMU (Inertial Measurement Unit) sensor was fixed on plastic pipes of different diameters and lengths. Vibration data collected in different scenarios (pressure, etc. factors) were transferred to the ThingSpeak platform over the internet. The transferred data could be monitored in real-time on a server. Physical damage in the pipes was detected using signal pre-processing, feature extraction, and feature selection algorithms on vibration data. In the study, damages were classified using machine learning-based classification (Decision Trees, k-Nearest Neighbors, Linear Discriminant, Support Vector Machines) methods to predict the type of damage (solid, hole, multi-hole). The data set revealed within the scope of the study is thought to lead to scientific studies in this field. The results obtained are close to the state-of-the-art results.

Machine Learning-based Classification of Developing Countries and Exploration of Country-Specific ODA Strategies

Introduction: This study aimed to develop a systematic methodology for classifying recipient countries using machine learning, with the premise that tailoring mid- to long-term ODA strategies to country characteristics is essential. Additionally, it sought to propose ODA policy directions considering the unique attributes of classified developing countries. Methods: The research analyzed 166 countries, including both developed and developing nations, using SDG scores and GDP per capita as key indicators. Machine learning techniques, specifically neural network analysis and decision tree analysis, were employed for classification. Results: The analysis resulted in the classification of the 166 countries into 12 distinct groups, with seven nodes representing developing countries. Each group exhibited unique characteristics that informed the development of country-specific ODA strategies Conclusions: This study successfully developed a systematic classification methodology for recipient countries using machine learning. The resulting classification and proposed ODA strategies for each group provide a foundation for more targeted and effective ODA policies. This approach enables policymakers to tailor their strategies to the specific needs and characteristics of different developing country groups, potentially improving the impact and efficiency of ODA efforts.

Genome-Wide DNA Methylation Identifies Potential Disease-Specific Biomarkers and Pathophysiologic Mechanisms in Irritable Bowel Syndrome, Inflammatory Bowel Disease, and Celiac Disease.

Irritable bowel syndrome (IBS), inflammatory bowel disease (IBD), and celiac disease (CeD) present with similar gastrointestinal (GI) symptoms. DNA methylation-based biomarkers have not been investigated as diagnostic biomarkers to classify these disorders. We aimed to study DNA methylation profiles of IBS, IBD, CeD, and healthy controls (HC), develop machine learning-based classifiers, and identify associated gene ontology (GO) terms. Genome-wide DNA methylation of peripheral blood mononuclear cells from 315 patients with IBS, IBD, CeD, and HC was measured using Illumina's 450K or EPIC arrays. A methylation dataset on 304 IBD and HC samples was used for external validation. Differential methylation was measured using general linear models. Classifiers were developed using penalized generalized linear models using double cross-validation controlling for confounders. Functional enrichment was assessed using GO. Three hundred and fifteen participants (148 IBS, 47 IBD, 34 CeD, and 86 HC) had DNA methylation data. IBS-IBD and IBD-CeD showed the highest number of differentially methylated CpG sites followed by IBD-HC, CeD-HC, and IBS-HC. IBS-associated genes were enriched in cell adhesion and neuronal pathways, while IBD- and CeD-associated markers were enriched in inflammation and MHC class II pathways, respectively (p < 0.05). Classification performances assessed using area under the receiver operating characteristic curves (AUC) for IBS-IBD, IBS-CeD, and IBD-CeD were 0.80 (95% CI = 0.7-0.87, p = 6.75E-10), 0.78 (95% CI = 0.68-0.86, p = 4.57E-10), and 0.73 (95% CI = 0.62-0.83, p = 0.03), respectively. The performance of IBD-HC was successfully validated using external data (AUC = 0.74 [95% CI = 68-0.80, p < 0.001]). Blood-based DNA methylation biomarkers can potentially distinguish chronic GI disorders that present with similar symptoms. GO suggested functional significance of the classifiers in disease-specific pathology.

Non-invasive multi-cancer detection using DNA hypomethylation of LINE-1 retrotransposons.

The detection of circulating tumor DNA, which allows non-invasive tumor molecular profiling and disease follow-up, promises optimal and individualized management of patients with cancer. However, detecting small fractions of tumor DNA released when the tumor burden is reduced remains a challenge. We implemented a new highly sensitive strategy to detect base-pair resolution methylation patterns from plasma DNA and assessed the potential of hypomethylation of LINE-1 retrotransposons as a non-invasive multi-cancer detection biomarker. The DIAMOND (Detection of Long Interspersed Nuclear Element Altered Methylation ON plasma DNA) method targets 30-40,000 young L1 scattered throughout the genome, covering about 100,000 CpG sites and is based on a reference-free analysis pipeline. Resulting machine learning-based classifiers showed powerful correct classification rates discriminating healthy and tumor plasmas from 6 types of cancers (colorectal, breast, lung, ovarian, gastric cancers and uveal melanoma including localized stages) in two independent cohorts (AUC = 88% to 100%, N = 747). DIAMOND can also be used to perform copy number alterations (CNA) analysis which improves cancer detection. This should lead to the development of more efficient non-invasive diagnostic tests adapted to all cancer patients, based on the universality of these factors.

Ensemble and Non-Ensemble Machine Learning-Based Classification of Liver Cirrhosis Stages

Cirrhosis is a chronic liver condition characterized by gradual scarring of the tissue in the liver, which then leads to one of the more serious health problems. Early diagnosis and detection of this condition are critical to managing the patient's situation and planning his treatment. Machine learning is a computer science field in which many complex issues have otherwise been successfully resolved, especially in medicine. This work focuses on constructing an artificial intelligence system, assisted by machine learning algorithms, to help professionals diagnose liver cirrhosis at its early stage. In this paper, four different models have been constructed with the aid of clinical parameters of patients and machine learning techniques: Random Forest, KNN, histogram-based Gradient Boosting, and Soft Voting. Two Feature selection methods (Chi-Square and mutual information) have been combined to select the most relevant features in the dataset. Then non-ensemble and ensemble methods are used to detect the condition. The random forest model achieved the highest score among other model with 97.4 % accuracy with a 10-fold Cross-validation method.

Machine learning-motivated trace triethylamine identification by bismuth vanadate/tungsten oxide heterostructures.

Triethylamine, an extensively used material in industrial organic synthesis, is hazardous to the human respiratory and nervous systems, but its accurate detection and prediction has been a long-standing challenge. Herein, a machine learning-motivated chemiresistive sensor that can predict ppm-level triethylamine is designed. The zero-dimensional (0D) bismuth vanadate (BiVO4) nanoparticles were anchored on the surface of three-dimensional (3D) tungsten oxide (WO3) architectures to form hierarchical BiVO4/WO3 heterostructures, which demonstrates remarkable triethylamine-sensing performance such as high response of 21 (4 times higher than pristine WO3) at optimal temperature of 190°C, low detection limit of 57ppb, long-term stability, reproducibility and good anti-interference property. Furthermore, an intelligent framework with good visibility was developed to identify ppm-level triethylamine and predict its definite concentration. Using feature parameters extracted from the sensor responses, the machine learning-based classifier provides a decision boundary with 92.3% accuracy, and the prediction of unknown gas concentration was successfully achieved by linear regression model after training a series of as-known concentrations. This work not only provides a fundamental understanding of BiVO4-based heterostructures in gas sensors but also offers an intelligent strategy to identify and predict trace triethylamine under an interfering atmosphere.

Rapid mapping of land cover changes in tropical watershed in Indonesia using Google Earth Engine

Abstract Besitang watershed is one of the tropical watersheds in Indonesia. In the upstream area, Besitang watershed is part of the Gunung Leuser National Park, a conservation forest abundant with diverse wildlife, including primates and terrestrial mammals. Studying land cover changes in Besitang watershed is crucial. A challenge in studying land cover changes in tropical watersheds is cloud cover, thus necessitating land cover mapping studies in Besitang watershed using Google Earth Engine (GEE). GEE is a platform for processing geo-big data, providing extensive and cloud-free satellite image data. Therefore, this study aims to map land cover in Besitang watershed using machine learning-based classification on GEE. The land cover mapping process in Besitang watershed utilizes Landsat 8 satellite imagery. The input data includes bands 1 to 7 of Landsat 8 imagery and image transformations such as NDVI, NDWI, NDBI, BSI, EVI, NDTI, SATVI. The years selected for analysis are 2001 and 2021, with machine learning techniques tested including CART and Random Forest (RF). The results of this study indicate that RF is the best machine learning method for mapping land cover in Besitang watershed, using an image combination consisting of Band 2, Band 3, Band 4, Band 5, Band 6, Band, NDVI, NDWI, NDBI, BSI. The mapping results show a forest area change of approximately 352326.29 hectares, with the majority of changes to plantation.

Contribution of Scalp Regions to Machine Learning-Based Classification of Dementia Utilizing Resting-State qEEG Signals.

This study aims to investigate using eyes-open (EO) and eyes-closed (EC) resting-state EEG data to diagnose cognitive impairment using machine learning methods, enhancing timely intervention and cost-effectiveness in dementia research. A total of 890 participants aged 40-90 were included in the study, comprising 269healthy controls (HC), 356 individuals with mild cognitive impairment (MCI), and 265 with Alzheimer's disease (AD) from a cohort study. Resting-state EEG (rEEG) signals were recorded and transformed into relative power spectral density (PSD) data for analysis. The processed PSD data, representing 19 scalp regions, were then input into a Random Forest (RF) machine learning classifier to identify distinctive EEG patterns across the groups. Statistical comparisons between the groups were conducted using one-way ANOVA, applied to the relative PSD features extracted from the EEG data, to assess significant differences in EEG activity across the diagnostic categories. The study found that rEEG-based categorization effectively differentiates between cognitively impaired individuals and healthy individuals. The EO rEEG achieved the highest performance metrics across various models. For HC vs MCI (combined hemisphere), the accuracy, sensitivity, specificity, and AUC were 92%, 99%, 83%, and 96%, respectively. For HC vs AD (parietal, temporal, occipital), these metrics were 95%, 96%, 94%, and 99%. The HC vs CASE (MCI + AD) (combined hemisphere) results were 90%, 99%, 73%, and 92%. The metrics for HC vs MCI vs AD (frontal, parietal, temporal) were 89%, 88%, 94%, and 96%. The study demonstrates that EO rEEG can effectively distinguish between cognitive impairment and healthy states, leading to early diagnosis, cost-effective treatment, and better clinical outcomes for dementia patients. EO and EC rEEG models trained with relative PSD, particularly from parietal, temporal, occipital, and central scalp regions, can significantly assist clinicians in practice.

Distinguishing among standing postures with machine learning-based classification algorithms.

The purpose of our study was to evaluate the accuracy with which classification algorithms could distinguish among standing postures based on center-of-pressure (CoP) trajectories. We performed a secondary analysis of published data from three studies: Study A) assessment of balance control on firm or foam surfaces with eyes-open or closed, Study B) quantification of postural sway in forward-backward and side-to-side directions during four standing-balance tasks that differed in difficulty, and Study C) an evaluation of the impact of two modes of transcutaneous electrical nerve stimulation on balance control in older adults. Three classification algorithms (decision tree, random forest, and k-nearest neighbor) were used to classify standing postures based on the extracted features from CoP trajectories in both the time and time-frequency domains. Such classifications enable the identification of differences and similarities in control strategy. Our results, especially those involving time-frequency features, demonstrated that distinct CoP trajectories could be identified from the extracted features in all conditions and postures in each study. Although the overall classification accuracy was similar using time-frequency features (~ 86%) for the three studies, there were substantial differences in accuracy across conditions and postures in Studies A and B but not in Study C. Nonetheless, the models were far superior to the published results with conventional metrics in distinguishing between the conditions and postures. Moreover, a Shapley Additive exPlanation analysis was able to identify the most important features that contributed to the classification performance of the models.

Prediction of cardiovascular events after carotid endarterectomy using pathological images and clinical data.

Carotid endarterectomy (CEA) is a surgical treatment for carotid artery stenosis. After CEA, some patients experience cardiovascular events (myocardial infarction, stroke, etc.); however, the prognostic factor has yet to be revealed. Therefore, this study explores the predictive factors in pathological images and predicts cardiovascular events within one year after CEA using pathological images of carotid plaques and patients' clinical data. This paper proposes a two-step method to predict the prognosis of CEA patients. The proposed method first computes the pathological risk score using an anomaly detection model trained using pathological images of patients without cardiovascular events. By concatenating the obtained image-based risk score with a patient's clinical data, a statistical machine learning-based classifier predicts the patient's prognosis. We evaluate the proposed method on a dataset containing 120 patients without cardiovascular events and 21 patients with events. The combination of autoencoder as the anomaly detection model and XGBoost as the classification model obtained the best results: area under the receiver operating characteristic curve, accuracy, sensitivity, specificity, and F1-score were 81.9%, 84.1%, 79.1%, 86.3%, and 76.6%, respectively. These values were superior to those obtained using pathological images or clinical data alone. We showed the feasibility of predicting CEA patient's long-term prognosis using pathological images and clinical data. Our results revealed some histopathological features related to cardiovascular events: plaque hemorrhage (thrombus), lymphocytic infiltration, and hemosiderin deposition, which will contribute to developing preventive treatment methods for plaque development and progression.

Classification of IPB variety of cayenne pepper genotypes using physical characteristics during the growing period until harvest using machine learning

In agrarian sectors, the prevalence of adulteration from seed selection to final agricultural outputs is a significant concern. This deceptive practice not only spans across distinct plant species but also extends to varieties within the same species, as observed in chili cultivation. Addressing this issue, the current study aims to develop a machine learning-based classification model that discriminates between cayenne pepper genotypes of IPB varieties and others. The model harnesses a range of plant physiological parameters monitored from the growth phase through to harvest. These parameters include plant and dichotomous heights, age at flowering, age at harvest, fruit stalk length, chili dimensions, individual chili weight, total number of chilies per plant, and overall yield productivity. Employing a dataset of 45 plant samples—30 from IPB varieties and 15 from non-IPB varieties—this research evaluates four machine learning algorithms: linear discriminant analysis (LDA), k-nearest neighbors (kNN), decision tree (DT), and random forest (RF). To ascertain the robustness of the proposed model, the study also investigates the impact of varying data split ratios, including 90/10, 75/25, and 50/50, on model performance. Preliminary results indicate that the DT classifier, with an accuracy exceeding 80%, successfully differentiates the IPB cayenne pepper genotypes from other varieties. This promising conceptual model warrants further validation and enhancement through future research on a larger dataset, paving the way for its practical application in ensuring the authenticity of agricultural produce.

Machine learning-based classification for Single Photon Space Debris Light Curves

The growing number of man-made debris in Earth’s orbit poses a threat to active satellite missions due to the risk of collision. Characterizing unknown debris is, therefore, of high interest. Light Curves (LCs) are temporal variations of object brightness and have been shown to contain information such as shape, attitude, and rotational state. Since 2015, the Satellite Laser Ranging (SLR) group of Space Research Institute (IWF) Graz has been building a space debris LC catalogue. The LCs are captured on a Single Photon basis, which sets them apart from CCD-based measurements. In recent years, Machine Learning (ML) models have emerged as a viable technique for analysing LCs. This work aims to classify Single Photon Space Debris LCs using the ML framework. We have explored LC classification using k-Nearest Neighbour (k-NN), Random Forest (RDF), XGBoost (XGB), and Convolutional Neural Network (CNN) classifiers in order to assess the difference in performance between traditional and deep models. Instead of performing classification on the direct LC data, we extracted features from the data first using an automated pipeline. We apply our models on three tasks, which are classifying individual objects, objects grouped into families according to origin, and grouping into general types. We successfully classified Space Debris LCs captured on Single Photon basis, obtaining accuracies as high as 90.7%. Further, our experiments show that the classifiers provide better classification accuracy with automated extracted features than other methods.

Estimating afforestation related forest cover change using data fusion and machine learning

Abstract The rapid increase in population and changes in land use have led to the issue of climate change, which is threatening the overall human well-being in general, and particularly the forest resources. Recognizing the rapid decline in the forest cover and in adherence to the Bonn Challenge, Pakistan has initiated the Billion Tree Afforestation Project (BTAP) to restore forests. Hence, there is a need to analyze the spatio-temporal dynamics of forest cover to assess the efficacy of BTAP. The objectives of this study were: (1) to develop machine learning methods that combine Sentinel-1 and Sentinel-2 data to characterize forest cover; and (2) to characterize the forest cover dynamics in the study area during the study period. In the study area, the land cover was classified using two machine learning models: random forests (RF) and support vector machines (SVM). We then used the models to create forest cover maps for the period of 2016 to 2022. Based on the classifications of land cover, the study area was classified into forest and non-forest classes. Finally, the spatiotemporal distribution of the changes induced by afforestation was generated. The results demonstrate an increase of 3.7% in forest cover in the study area during the study period. The increase in forest cover was more prominent in the northern and central regions as compared to that of the southern region. In terms of species, the increase in broadleaved forests was more prominent. The results show that RF produces superior results as compared to the SVM, with overall accuracy and kappa coefficient of 94%–97% &amp; 0.93–0.96 respectively. The overall accuracy and Kappa coefficient of the SVM model range from 92%–94% &amp; 0.91–0.95. The techniques used in this study are cost-effective for accurately monitoring changes in forest cover.