Journal of Innovations in Data Science and Big Data Management

An Intelligent Machine Learning Framework for Early Cardiovascular Disease Risk Prediction

Gagana G — 2026-09-09

Heart disease continues to be a major health problem across the world, and identifying it at an early stage can help improve treatment and reduce serious complications. In recent years, machine learning has become a useful approach for studying patient health records and predicting the possibility of heart disease. This survey reviews and compares different machine learning techniques such as Logistic Regression, Decision Tree, Support Vector Machine, Random Forest, K-Nearest Neighbors, Gradient Boosting, and XGBoost. It also looks at common data preparation methods, including data cleaning, feature selection, handling imbalanced datasets using SMOTE, and model tuning to improve prediction results. From the studies reviewed, ensemble-based methods like XGBoost and Gradient Boosting generally provide more consistent and accurate predictions than many traditional algorithms. This also discusses the advantages, limitations, and research opportunities in this area. Overall, the survey shows that machine learning can support doctors by providing faster and more reliable predictions, making early diagnosis of heart disease more effective.

AI Methods and Deployment Challenges from Prediction to Trusted Operation in Renewable-Rich Smart Grids: A Review

Ganesh Kumar Gautam — 2026-09-28

The inclusion of renewable energy sources, distributed energy resources, and electric vehicles in the power grid has transformed conventional power grids into complex, multi-directional cyber-physical systems, for which traditional model-based tools are inadequate and/or linear. Artificial Intelligence (AI) and Machine Learning (ML) have therefore become key enablers of next-generation smart grids, with AI-driven data-based solutions helping forecast energy loads and renewable generation, diagnose faults and protect against them, optimise microgrids and electric-vehicle charging, and support cybersecurity and explainable decision-making. This work summarizes the current AI/ML applications for smart grid optimization and renewable-energy integration, based mainly on literature published from 2024 to 2026, including peer-reviewed studies and recent preprints. It presents recent breakthroughs in deep learning-based forecasting, reinforcement learning-based microgrid energy management, and digital-twin-based grid monitoring and AI-based cyber defense, and places these advances in the context of the fast-digitalizing power system in India as an illustrative example from emerging economies. There are persistent data quality issues, model interpretability concerns, real-time scalability issues, and AI-specific security concerns, which are identified, and directions for future research are proposed. The review aims to serve as a single reference to help researchers, utility engineers, and policymakers design "smart" power systems that are able to be powered by renewables.

Predictive Modeling in Healthcare using Deep Learning: A Comprehensive Review

Chandrasekhar Koppireddy — 2026-10-06

Predictive modelling is essential to the healthcare sector for enhancing clinical outcomes and to maximize scarce resources in the healthcare system. The increasing digitization of healthcare systems creates a vast amount of data, including electronic health records (EHRs), medical imaging, genome data, and data from wearable devices. Traditional analytical techniques frequently cannot be applied to the analysis of this high-dimensional and heterogeneous data. Many of these challenges are addressed by a subfield of artificial intelligence (AI) known as deep learning, which constructs hierarchical representations directly from raw data, without relying on much feature engineering. This study offers a systematic review of predictive modeling in healthcare with deep learning, which includes popular architectures like Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM), Transformer-based systems, and hybrid multi-modal systems. In addition to summarizing previous research, this review critically compares these architectures in terms of their strengths and weaknesses, data needs, and application to specific prediction tasks in healthcare. Applications are explained for disease prediction, medical imaging, patient monitoring in ICUs, personalized medicine, and prediction of health in populations. The study also provides a detailed discussion on key open problems such as interpretability, data quality, algorithmic fairness, regulatory compliance, and clinical integration, as well as new trends, such as federated learning, explainable AI, foundation models, Vision Transformers, and Retrieval-Augmented Generation (RAG) in healthcare. In conclusion, the evidence suggests that deep learning has significant promise of changing the landscape of healthcare delivery, but achieving this will involve continued collaboration among many disciplines to build trust, transparency, and safe clinical deployment.