Healthy Lifestyle Pattern Classification Using the XGBoost Algorithm on the Lifestyle and Wellbeing Dataset

Abstract

Using the Extreme Gradient Boosting (XGBoost) algorithm with the Lifestyle and Wellbeing dataset from the Kaggle platform, this study attempts to categorize healthy lifestyle patterns. The growing incidence of undesirable behaviors that have a direct impact on people's health and well-being, like stress, inactivity, and a poor work-life balance, is what spurred the research. Using 500 randomly chosen data samples that cover a range of behavioral and psychological lifestyle variables, the study uses a quantitative computational method. Data cleaning, standardization, and target label generation using a health index, which is computed as the mean of positive variables less the mean of negative variables, are all included in the preprocessing stage. With n_estimators = 100, learning_rate = 0.1, and max_depth = 4, the XGBoost model was trained using Python in the Google Colab environment. According to the results, the model's accuracy was 97%, and its precision, recall, and F1-score were all balanced. SLEEP_HOURS, DAILY_STRESS, and WORK_LIFE_BALANCE_SCORE are the most significant factors in predicting a healthy lifestyle, suggesting that psychological stability and sufficient rest are important factors in determining general well-being. As a basis for creating adaptive healthy lifestyle recommendation systems and future research incorporating physiological data from wearable devices to improve prediction accuracy, the study concludes that XGBoost successfully classifies lifestyle patterns and offers comprehensible insights into behavioral factors that contribute to health.