Stress Prediction in Modern Life Using Machine Learning Algorithms

Abstract

The healthcare sector is rapidly advancing with the integration of innovative technologies such as machine learning (ML). Connected and smart healthcare, particularly through devices like smartwatches and wearable sensors, offers a wide range of benefits. These advancements not only enhance the quality of care but also improve the patient experience and the overall efficiency of healthcare systems. Data on our physical and mental health can be harnessed to transform medical decision-making, support research into new treatments, and deepen our understanding of complex diseases. However, traditional ML systems often face limitations, particularly in real-time processing and resource optimization, limiting their application in critical situations. This paper aims to optimize data analysis to tackle a significant health concern: stress. By applying various machine learning techniques, it proposes the development of an innovative system capable of real-time and remote health monitoring. The system strives to offer personalized, efficient solutions for detecting, analyzing, and predicting stress levels, ensuring that care is customized to each individual's needs. This study underscores the potential of feature selection algorithms, such as Random Forest (RF), k-Nearest Neighbour (KNN), Gradient Boosting, Linear Discriminant Analysis (LDA), and Decision Tree (DT),in reducing dimensionality. The application of these algorithms resulted in improved classification outcomes.