Performance Analysis Of Machine Learning Algorithms To Foresee The Psychological Wellbeing Of An Individual

Abstract

In Healthcare systems, for prediction psychological health several attributes have been developed. These psychological attributes integrate the study of outcomes of attributers such as stress, anger, fatigue and depression. The existing models are built up on multiple regression method and theories give a reason to effect model of the psychological health prediction. Though, this raised algorithm depends on the outcomes of the survey. The association among these inspection replies and psychological attributes are not linear due unevenness of data. For inconsistency in the data, the dependability of the algorithm cuts and finally maintenance cost of algorithm revision rises. Similarly, when new attributes are measured, the whole algorithm should be rebuilt from the beginning. Our effort is to observes the probability of machine learning algorithms to foresee the psychological attributes grounded on the reconstructed responses. The paper compares 4 machine learning algorithms i.e. Multi-layer perceptron, Principal component analysis, Random forest, and k nearest neighbour regression, are related and the experimental results are obtained.