A machine learning aid to predict diseases based on lifestyle and symptoms

Abstract

Currently with this pandemic situation, health is given topmost priority. Lifestyle and diet are the two foremost features that are measured to effect numerous illnesses. Sicknesses are mostly triggered by grouping of alteration, lifestyle choices and environments [4]. Taking precaution, early detection of diseases and awareness about possible health breakdown can create a drastic change which can eventually lead to a good health. This study aims to predict lifestyle diseases an individual is susceptive to and create awareness about healthy life style.

A huge dataset consisting of patient attributes like unhealthy eating habits, obesity, alcoholic, stress and many other parameters are taken as input and would be modelled thereby classifying a lifestyle disorder. Based on the output, symptoms would be given as input and using Naïve Bayes having class label as Disease name and symptoms as parameters, the model would predict current or future occurrence of disorders. Based on symptoms and habits, the proposed system is capable of suggesting home remedies to control the disease from becoming worse until patient meets doctor.

 After applying the algorithm and building the model, lifestyle disorder prediction yielded 98% accuracy whereas symptom-based disorder prediction yielded 92.42 % accuracy along with respective home remedy suggestion.