AI-powered Predictive Model for Stroke and Diabetes Diagnostic


 
 
 
 Abstract
 The classification task known as MNIST digit recognition involves identifying handwritten numbers into their corresponding values. Although there are numerous approaches proposed for this type of task, they typically face issues in achieving high accuracy. One method that can improve single models' performance is through ensemble learning. The goal of this study is to explore the use of various learning techniques, such as boosting and bagging, in combination with random forest models and decision trees, to improve the performance of MNIST digit recognition with regard to accuracy. We then perform evaluations on these methods using various metrics, such as recall, precision, accuracy, and F1. The findings of this study provide valuable insight into the various advantages of ensemble methods for the MNIST digit recognition task. It also highlights the need to explore these techniques in the context of machine learning. The objective of this study is to investigate the use of ensembles in improving the accuracy of MNIST digit recognition. We performed evaluations on two popular methods, namely boosting and bagging, with random forest and decision tree models. The evaluation parameters included F1 score, recall, accuracy, and precision. The results of the evaluations revealed that both boosting and bagging methods performed well in terms of their evaluation metrics. In most cases, the decision tree performed better than the random forest. However, the random forest method was able to achieve the highest accuracy, which is 99 percent. The findings of the evaluation revealed that ensembles can help improve single models' accuracy in MNIST digit recognition. On the other hand, the random forest method is a promising option for this task. The exact results of the evaluations will vary depending on the evaluation and implementation metrics. More research is needed to confirm their generalizability. The study emphasizes the value of exploring ensembles in machine learning systems, as well as the potential advantages of performing MNIST digit recognition using them.
 
 
 

Cancer recurrences are poorly recorded within electronic health records around the world. This hinders research into the efficacy of cancer treatments. Currently, the retrospective identification of recurrence/progression diagnosis dates is achieved by staff who manually review patients' health records. This is expensive, time-consuming, and inefficient. Machine Learning models may expedite the review of health records and facilitate the assessment of alternative cancer therapies. This paper evaluates the use of four machine learning models (random forests, conditional inference trees, decision trees, and logistic regression) in identifying proxy dates of epithelial ovarian cancer recurrence/progression from chemotherapy data, in 531 patients at Leeds Teaching Hospital Trust. The random forest achieved the highest F1 score of 0.941 (95% confidence interval 0.916-0.968) when identifying recurrence events. Both the random forest and decision tree models' classifications closely conform to chart-reviewed time to next treatment, serving as a surrogate for recurrence-free survival. Additionally, all models reached an F1 score >0.940 when identifying patients whose cancer recurred/progressed. Our models proficiently identify both proxy dates for recurrence/progression diagnoses and patients whose cancer recurred/progressed. Considering the similar performance of the random forest and decision tree, model preference should be determined by the interpretability required to assist chart review and the ease of implementation into existing architecture.

<abstract><p>In the realm of machine learning, where data-driven insights guide decision-making, addressing the challenges posed by class imbalance in datasets has emerged as a crucial concern. The effectiveness of classification algorithms hinges not only on their intrinsic capabilities but also on their adaptability to uneven class distributions, a common issue encountered across diverse domains. This study delves into the intricate interplay between varying class imbalance levels and the performance of ten distinct classification models, unravelling the critical impact of this imbalance on the landscape of predictive analytics. Results showed that random forest (RF) and decision tree (DT) models outperformed others, exhibiting robustness to class imbalance. Logistic regression (LR), stochastic gradient descent classifier (SGDC) and naïve Bayes (NB) models struggled with imbalanced datasets. Adaptive boosting (ADA), gradient boosting (GB), extreme gradient boosting (XGB), light gradient boosting machine (LGBM), and k-nearest neighbour (kNN) models improved with balanced data. Adaptive synthetic sampling (ADASYN) yielded more reliable predictions than the under-sampling (UNDER) technique. This study provides insights for practitioners and researchers dealing with imbalanced datasets, guiding model selection and data balancing techniques. RF and DT models demonstrate superior performance, while LR, SGDC and NB models have limitations. By leveraging the strengths of RF and DT models and addressing class imbalance, classification performance in imbalanced datasets can be enhanced. This study enriches credit risk modelling literature by revealing how class imbalance impacts default probability estimation. The research deepens our understanding of class imbalance's critical role in predictive analytics. Serving as a roadmap for practitioners and researchers dealing with imbalanced data, the findings guide model selection and data balancing strategies, enhancing classification performance despite class imbalance.</p></abstract>

Quantitative Structure-Property Relationship (QSPR) models for Minimum Ignition Energy (MIE) prediction of combustible dusts using machine learning

-Stroke diagnosis is a time-critical medical task that requires rapid and precise identification to initiate appropriate treatment and prevent severe neurological damage. This study introduces a machine learning–driven diagnostic framework designed to identify stroke using neuroimaging data. A comprehensive set of machine learning models—Logistic Regression, Support Vector Machine (SVM), Random Forest, Decision Tree, and Convolutional Neural Network (CNN)—were trained and evaluated on a curated dataset of brain images to classify stroke and non-stroke cases. The proposed system leverages the strengths of both classical and deep learning algorithms. Logistic Regression enables examination of relationships between neuroimage features and stroke presence, while SVM effectively separates complex feature patterns. Random Forest and Decision Tree models provide structured rule-based classification derived from key imaging attributes. The inclusion of a CNN significantly enhances diagnostic accuracy by extracting high-level spatial and textural features from neuroimages without manual feature engineering. Experimental results indicate that the CNN model outperforms all other techniques, achieving 95.6% accuracy, 94.2% sensitivity, and 96.5% specificity. Random Forest and SVM also show strong performance with accuracies of 93.1% and 92.5%, respectively. The findings confirm the potential of machine learning, particularly deep learning, in improving stroke diagnosis and supporting clinicians with faster and more reliable decision-making. By harnessing modern computational intelligence, this framework contributes toward reducing diagnostic delays, enhancing patient outcomes, and lowering the overall healthcare burden associated with stroke. Key Words: Brain Imaging, Classification Algorithms, Convolutional Neural Network (CNN), CT Scan, Decision Tree, Deep Learning, Diagnostic Model, Logistic Regression, Machine Learning, MRI Scan, Neuroimages, Random Forest, Stroke Detection, Stroke Identification, Support Vector Machine (SVM).

The detection of cryptographic attacks is a vital aspect of maintaining cybersecurity, especially as digital infrastructures become increasingly intricate and susceptible to sophisticated threats. This systematic review aims to examine and compare a range of machine learning approaches applied to cryptographic attack detection, focusing on their performance in terms of detection rates, efficiency, and overall effectiveness. A comprehensive review and meta-analysis were conducted, focusing on existing research that utilized machine learning models for identifying cryptographic attacks. The models included in the review were Naïve Bayes, C4.5, Random Forest, Decision Tree, K-Means, and Particle Swarm Optimization (PSO) combined with Neural Networks. Studies were selected based on their relevance to cryptographic security, with particular attention paid to performance metrics like classification accuracy, precision, recall, and area under the curve (AUC). The findings indicated that the C4.5 decision tree model achieved a high classification rate of 98.8%, while both Random Forest and Decision Tree models performed with an accuracy of 99.9%, making them highly suitable for real-time attack detection. Additionally, the PSO + Neural Network model showed enhanced detection precision, illustrating the value of integrating optimization techniques with machine learning models. The use of machine learning, especially with ensemble methods such as Random Forest and Decision Trees, proves to be highly effective for cryptographic attack detection. The study underscores the necessity for customized machine learning solutions in cybersecurity, balancing both high accuracy and operational efficiency. Further research should focus on the real-world deployment of hybrid models to confirm their practical effectiveness.

Aim: The aim of the work is to evaluate the accuracy and precision in predicting heart disease using Support Vector Machine (SVM), Random forest (RF), Logistic Regression (LR), Decision Tree (DT) Classification algorithms. Materials and Methods: Classification algorithm is appealed on a heart dataset which consists of 180 records. A framework for heart disease prediction in the medical sector comparing Random forest, Logistic Regression, Decision Tree and SVM classifiers has been proposed and developed. The sample size was calculated as 55 in each group using G power 80%. Sample size was calculated using clincalc analysis, with alpha and beta values 0.05 and 0.5, 95% confidence, pretest power 80% and enrolment ratio 1. Results: The Novel Random Forest Algorithm (92.13%), Support Vector Machine (62.51%), Logistic Regression (84.89%), Decision Tree (86.25%) classifiers produce respectively. SVM, RF exists a statistically significant difference between the two groups (p=0.001,p=.004;p<0.05).LR, RF exists a statistically insignificant difference between the two groups (p=.103, P=.080;p>0.05) both with confidence interval 95%. Hence Random forest is better than SVM, RF, DT classifiers. Conclusion: The results show that the performance of RF is better when compared with SVM, LR and DT in terms of both precision and accuracy.

FakeNews is one of the most popular phenomena that have considerable effects on our social life, especially in the political domain. Nowadays, creating fake news becomes very easy because of users' widespread using the internet and social media. Therefore, the detection of elusiveness news is a crucial problem that needs to be considerable mainly because of its challenges like the limited amount of the benchmark datasets and the amount of the published news every second. This research proposed utilizing two different machine learning algorithms (random forest and decision tree (J48)) to detect the fake news. In this paper, the full dataset size equals 20,761 samples, while the testing sample size equals 4,345 samples.The preprocessing steps start with cleaning data by removing unnecessary special characters, numbers, English letters, and white spaces, and finally, removing stop words is implemented. After that, the most popular feature extraction method (TF-IDF) is used before applying the two suggested classification algorithms. The results show that the best accuracy achieved equals 89.11% using the decision tree model while using the random forest; the accuracy achieved equals 84.97 %.

In modern times cardio pulmonary diseases are increasing at an alarming rate. It is a range of serious disorders that affects both the heart (cardio) and the lungs (pulmonary) at times resulting in death. In spite of continuous develop-ment in the field of medical science – both in the surgical and oral drugs, death due to cardiac arrest has remained a major cause of headache for the global population. Research suggests that mortality due to cardiac failure is mainly occurring in the early morning due to high blood pressure fluctuations. An advanced technology like machine learning has shown a new horizon of hope in the medical arena. ML carefully helps to select the patterns best suitable for data sets with various types of algorithms along with different correlation techniques. There are many features that are responsible for the heart disease like age, BP, sodium, cretonne, ejection fraction etc. After identifying important features by applying feature selection and ranking techniques the selected features are used as an input for the different ML algorithms. After imple-menting Random Forest algorithm it was observed that two sets of data tree emerged as comparison of results. Firstly a Decision Tree Classifier was select-ed to control the entire data sets, as well as to check the results with the gen-erated tree for the selected samples. After generating selected samples its re-sults got compared with the classification graph of the decision tree to achieve better result. It was observed that ML algorithms like Random Forest and De-cision Tree Classifier Techniques can helps to predict the major reasons for cardio vascular diseases

Blood sugar issues are a major health issue worldwide, with their incidence growing rapidly and affecting human health, economic systems, and societal structures. If diabetes remains untreated and undiagnosed, it can cause blood sugar levels to vary significantly, potentially damaging essential organs like the kidneys, eyes, and arteries of the heart in critical cases. As a result, there is an increasing focus on the prevention and early detection of diabetes mellitus within the medical community. Utilizing machine learning algorithms to analyze appropriate datasets for early disease prediction could prove life-saving. The objective of this paper is to examine four algorithms that are proposed to enhance the diagnosis of diabetes. This research analyzes the effectiveness of various machine learning algorithms in processing datasets with minority classes. The evaluation was based on the classification report (including accuracy, precision, recall, and F1-score), the confusion matrix, and the ROC AUC. The Diabetes Prediction Dataset is used to evaluate four machine learning algorithms. The classifier that deserves a singular mention is the Artificial Neural Network (ANN), which achieves a 97% accuracy rate. This demonstrates its capability of classifying instances that are common and less common types. The Random Forest and Decision Tree models also perform well in terms of their ability to deliver strong performance, and the outcome shows some incremental differences, suggesting their ability to manage the dataset is quite high. However, the Support Vector Machine (SVM) model performs worse than all the above models at 96.36% and seems to struggle with the correct classification of less frequent instances. Therefore, it would be problematic to distinguish between classes that are prominent and those that are not. Notably, the ANN, Random Forest, and Decision Tree models effectively identify cases that are more likely to capture rare cases, an important aspect when dealing with datasets that have class imbalance.

Three machine learning classifiers: random forest, decision tree and support vector machine were used to build predictive models of an anti-mycobacterial ChEMBL database and evaluated for their predictive capability. Before the development of predictive models, data pre-processing was carried out to fix the class imbalance problem by applying cost-sensitive classifier, and filtration of data instance by supervised synthetic minority oversampling technique (SMOTE), spread subsample and resample method. The statistical evaluation indicated that random forest model was the best model as it showed the best accuracy 93.83%, specificity 90.5%, receiver operating characteristic (ROC) 0.984, MCC 0.772 and kappa statistics 0.768 in comparison to other models whereas LibSVM showed the highest sensitivity 94.4% compared with others. Additionally, toxicity predictive models based on SingleCellcall DSSTox carcinogenicity database (AID1189) was developed which resulted in random forest model as the best model. The deployment of both RF predictive models on two unknown datasets resulted in 1317 compounds out of 1554 approved drugs and 2234 compounds out of 18,746 ChEMBL anti-malarial dataset as non-toxic and anti-mycobacterial compounds. Thus machine learning models present highly efficient methods to find out novel hit anti-mycobacterial compounds. We suggest that such machine learning techniques could be very useful to screen drug candidates not only for tuberculosis but also for other diseases.

With the rapid development of the global economy, P2P loan mode is increasingly favored by SMEs and individuals. Due to the relatively small loan amounts and uncertain risks associated with these groups, the evaluation of multiple factors to assess the risk and determine whether to grant or successfully obtain a loan is of significant importance to both borrowers and lenders. It is also an essential component of modern financial risk assessment. Based on the P2P data of Prosper Funding LLC in the United States, this paper established a default prediction model using random forest, decision tree, Logistic, GBDT, XGboost, LGBM models. The results demonstrate that models such as random forest, decision tree, LGBM and other models exhibit good fit and predictive capabilitie. Additionally, we identified a set of highly correlated features with loan default, including: loan amount, whether to accept partial payment of loan amount, debt income ratio of lender, previous cumulative loan batches, expected return rate of lender and more. These findings will contribute to the platform's improved ability to identify valuable potential credit customers.

This article investigates the problem of classifying network devices based on their security parameters using machine learning. Due to the constant growth of threats in cyberspace and the need to ensure a high level of network security, the relevance of using machine learning technologies to identify and classify secure devices is exceptionally high. Therefore, the article considers the specifics of applying machine learning algorithms for classification and regression tasks in network environments. Particular attention is paid to a short review of the algorithms most commonly used for classification tasks. The work describes in detail the process of developing a machine-learning model aimed at classifying network devices according to their security indicators. It considers the selection of appropriate parameters for model training, the process of data preprocessing, and the selection and adjustment of the classification algorithm. The results of model training on actual data are also presented. The process of training and evaluating the accuracy and efficiency of the model is described. The results are analyzed, and the choice of optimal hyperparameters is justified. As a result of the study, an effective machine learning model has been developed that can accurately classify network devices by security level, improving network security when selecting potentially secure devices. The study found that the Random Forest and Decision Tree models showed the highest accuracy in predicting the security state of network devices compared to other models, such as Logistic Regression, k-NN, and Gradient Boosting. The performance of the trained model was tested on a validation dataset. The Decision Tree model correctly predicted the security level of approximately 78% of network devices.

The urban passenger transportation sector is a major energy consumer in Canada, accounting for nearly 48% of transport-related energy use and is heavily reliant on fossil fuel-based vehicles. Electrifying private vehicles offers a promising solution to reduce fossil fuel consumption and greenhouse gas (GHG) emissions, but its full potential remains underexplored. This study focuses on the Greater Toronto and Hamilton Area (GTHA). It examines how residential dwelling types influence daily household transportation energy demand, with a specific focus on electrification barriers in multi-unit residential buildings (MURBs). Using personal and household travel surveys, the study applies supervised machine learning models, such as Random Forest and Decision Tree models, to impute vehicle engine types and estimate daily transportation energy use. The results show significant variation in energy demand by dwelling type, with detached homes consuming the most, followed by MURBs and townhouses. Moreover, there is an increase in reliance on private vehicles and ridesharing post-pandemic. Scenario analysis reveals that a complete transition to electric vehicles (EVs) has the potential to reduce daily household private vehicle energy consumption by up to 77.3%. The reductions vary by dwelling type, with detached homes projected to achieve a 54.4% decrease, while MURBs are expected to see a 14.8% reduction. Peak hour charging demand in 100% EV scenarios would reach 6640 GJ for houses and 1792 GJ for MURBs. These findings underscore the need for targeted policies to promote EV adoption, particularly for MURBs, and tailored incentives for households with detached homes. • Transportation energy demand varies by dwelling type, with detached homes highest. • Daily energy demand for private vehicles and ridesharing increased post-pandemic. • 100% EVs can reduce daily private vehicle energy consumption by 77.3%. • Detached homes may achieve a 54.4% reduction in energy demand, MURBs 14.8%. • Peak hour charging demand: 6640 GJ for detached homes, 1792 GJ for MURBs.

Stock Classification Prediction Based on Spark