Exploring the Potential of Royal Animal Optimization Algorithms for Diabetes prediction in Indian population data

Introduction: Diabetes remains a major public health concern, particularly in India, where its prevalence is rising due to lifestyle and genetic factors. Advanced AI-driven optimization techniques can enhance predictive models, improving early diagnosis and intervention strategies. Objectives: The objective of this study to explore the potential of Royal Animal Optimization (RAO) Algorithms, including the Lion Optimization Algorithm (LOA), Cheetah Optimization Algorithm (COA), and Polar Bear Optimization Algorithm (PBOA), to enhance predictive performance of ensemble algorithms. The prediction outcome of ensemble model is further explained with explainable artificial intelligence (XAI) model. Methods: Data Collection: The life style related data was collected from north east part of Rajasthan between July 2024 to December 2024. Feature Selection: The RAO algorithms are compared with evolutionary Optimization (EO) and Swarm Intelligence (SI) based algorithms by selecting the most relevant features, with logistic regression (LR) as the objective function. Predictive Model: An XGBoost classifier was implemented to perform diabetes classification. The RAO-based feature selection were integrated to optimize the model for high accuracy. Performance Metrics: The model's performance was evaluated using classification accuracy to ensure reliability and precision in diabetes risk assessment. Results: The XGBoost ensemble classifier along with ROA demonstrated superior performance, achieving a classification accuracy of 97.91%. This high-performance metric underscores the effectiveness of the proposed approach in identifying diabetes risk factors and predicting diabetes in the target population. Conclusions: This study demonstrates that the outcome of XAI model depressed, severely anxious and aged people found at high risk of diabetes. RAO outperformed than other optimization techniques such as, evolutionary and swarm optimization. The findings emphasize the role of AI-driven optimization techniques in addressing the growing burden of diabetes and other lifestyle-related diseases.