Medical Recommendation System for Personalized Disease Prediction

The use of machine learning (ML) and artificial intelligence (AI) have sped up the development of solutions in healthcare. Thesesolutions can move from static rule-based systems to adaptive, data-driven solutions and allow for faster, more intelligent diagnoses,prognostics, and treatment plans. In this study, we provide a more sophisticated, complete end-to-end medical recommendationsystem and the potential to not only predict diseases based on user-reported symptoms, but also recommend tailored advice onmedication prescriptions, diets, exercise plans, and other relevant actions. Essentially, the system uses an assortment of supervisedmachine learning algorithms to optimize disease diagnosis using Naïve Bayes, Decision Tree, Random Forest, and Support VectorMachine (SVM) models. However, our software takes the process further since we provide actionable recommendations in additionto predictions. In order for patients to access effective treatment options when the first-line medications are not suitable oravailable, we employ cosine similarity algorithms to pair any medication with alternatives. Along with prescribing medications, thesystem offers individualized diets and exercise programs based on each projected disease. This holistic approach considers bothshort-term treatment needs and long-term health goals to help people overcome nutritional deficits and chronic symptoms andestablish healthy habits that last. Our frame of reference is shifting to a holistic problem-solving model as compared to earlierpredictive models that simply worked to identify sickness. The capabilities are available through a scalable and easy-to-use webinterface built in Flask. The seamless transition between managing medications, lifestyle coaching, and predicting disease suggeststhat AI-powered platforms can formulate from lab prototypes into functional tools to manage health on a daily basis.