HEALTH GUARD-A Multiple Disease Prediction Model Based on Machine learning

Abstract

The aim of the paper is to present a new approach
 to predicting multiple lifestyle diseases such as diabetes and
 heart disease using machine learning techniques. The proposed
 approach is based on ensemble learning, genetic algorithm
based recursive feature elimination, and AdaBoost. The data is
 preprocessed using the Multiple Imputation by Chained Equa
tions (MICE) technique to handle missing data. This technique
 is used to impute missing values in the dataset by creating
 multiple imputations and then combining them to create a final
 dataset. The proposed approach also uses genetic algorithm-based
 recursive feature elimination to determine the optimal feature
 subset. This technique uses a genetic algorithm to iteratively
 remove features from the dataset until the optimal subset is
 found. The AdaBoost classification model is trained alongside
 other predictive models for multi-disease prediction. AdaBoost is
 an ensemble learning technique that combines mul- tiple weak
 classifiers to create a strong classifier. An extensive comparative
 study has been conducted to evaluate the effectiveness of the
 proposed model. The results show that the proposed methodology
 outperforms existing works in terms of prediction accuracy,
 precision, and recall. Overall, this study demonstrates the effec
tiveness of ensemble learning and genetic algorithm-based feature
 selection in predicting multiple diseases. The proposed approach
 has the potential to improve disease prediction accuracy and help
 healthcare professionals make more informed decisions.