Journal of Artificial Intelligence and Metaheuristics

In this review paper, the authors discuss the development and application of methods for modeling and control and comparison of viral spreading in society with fractional-order and ML techniques for data analysis. Some of the most well-known epidemiological models are based on traditional approaches to describing disease diffusion and often need to be more sufficient when mapping the realistic disease distribution. However, fractional-order models give more flexibility and accuracy due to the memory incorporated and interaction factors. Moreover, the amalgamation of ML and artificial intelligence allows the analysis of considerable and heterogeneous amounts of data, enabling real-time prediction and favorable outbreak response measures. This paper outlines some benefits of integrating these sophisticated techniques while discussing issues such as the quality of inputs, problems in the methods deployed, and issues of visibility of the methods deployed. Finally, it proposes better epidemic preparedness and response through interdisciplinary approaches that emphasize the role of these technologies in a society that is more vulnerable to epidemic diseases.

Influenza is often associated with millions of cases and hundreds of thousands of deaths each year, thus constituting a serious threat to public health. Traditional surveillance techniques employed in epidemiology are limited in forecasting impending outbreaks as caused by delays in receiving the relevant information and the dynamic nature of political environments. This review focuses on the available literature on the use of machine learning (ML) techniques in understanding and controlling influenza with an accent on all the sources of information available, including clinical papers, social networking sites and others. Applicable practices in classifying predictive modeling techniques, including deep learning and others, ensemble techniques, time series analysis, etc., have increased the speed and precision of the earlier results. Even so, the achievements made so far have not come on a silver platter as there are challenges, but not limited to data issues, model explain ability and strict validation processes. Some research areas are enhancing the present models to accommodate diverse virulent strains of the viruses and advancing extensive data analysis methods. It is noted in this review that machine learning strategies are essential in combating health issues and, thus, why such technologies can be deployed within a concise duration in the context of influenza epidemics for effective forecasting and resource management to salvage lives.

Heart attacks, or myocardial infarctions, are a primary cause of mortality worldwide, underscoring the importance of early and accurate diagnosis to improve patient outcomes. This paper reviews various Artificial Intelligence (AI) and Machine Learning (ML) techniques for heart attack diagnosis, focusing on both traditional algorithms and more complex models. The traditional algorithms are Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Logistic Regression (LR), and Decision Trees (DT). More complex models are Convolutional Neural Networks (CNN), Extreme Gradient Boosting (XGBoost), Auto-encoders, Artificial Neural Networks (ANN), and TSK Fuzzy Inference System (TANFIS). Additionally, the integration of optimization techniques, including the Grey Wolf Optimizer (GWO), Particle Swarm Optimization (PSO), and Jellyfish Optimization Algorithm (JOA) is explored to enhance model accuracy by selecting the most important features. Our findings indicate that ensemble and hybrid models, which combine ML with metaheuristic optimization, show significant potential in improving diagnostic performance and reducing overfitting. However, challenges remain, particularly regarding computational complexity and interpretability. This study provides insights into the strengths and limitations of different AI-based diagnostic models, contributing to the advancement of automated heart disease prediction systems.

This study pursues machine learning models for the task of smart homes' energy management with the use of a dataset that combines smart meter readings and weather conditions at the same time. The assessment of the Baseline Qualification and ARIMA models is done using various criteria, such as MSE, RMSE, and others. Most telling, the best performance is shown by ARIMA, which gets the lowest MSE score, 0.0693, in this instance. They show that such a model is optimal in forecasting energy consumption dynamics, and while they could be better, weather information helps improve the accuracy of the forecasts. The conduct helps uncover priceless information, allowing for the development of new smart home operating systems with a prospect of energy efficiency enhancement as well as a sustainable environment.

This paper proposes to evaluate how different machine learning techniques can be used to predict daytime power generation based on the "Daily Power Generation Data" data set. As a result of six models, which contain Random Forest Regressor, Decision Tree Regressor, Nearest Neighbors, Linear Regression, MLP Regressor, and SVR, a clear understanding has been accomplished by assessing the performance using multiple metrics. First, the Random Forest Regressor turned out to be the best in terms of the Mean Squared Error (MSE) of 3.57E-06, which was the lowest among the three ML models. The introduction of the paper highlights the role of precise planning of the power market and the consecutive sections describing the topic mathematically. The table below, with a total list of performance issues, explains why the Random Forest Regressor is the superior full-proof model using the lowest MSE, highest explained variance, and great resistance to outlying samples. The paper thus gave various useful approval criteria that we can largely choose the best model out of them because the Random Forest Regressor was able to get the highest performance metrics.