Symptom-Based Disease Prediction Using Machine Learning

Accurate prediction of disease is a major challenge in the current healthcare sector, where on-time and correct diagnosis can remarkably impact treatment outcomes. This research introduces a user-focused approach to predicting diseases, using machine learning along with a mobile application interface. Empowering individuals with the ability to assess their health and make well-informed decisions is the primary aim of this research. Our machine-learning model demonstrates exceptional accuracy, often achieving a 100% accuracy rate. This includes the incorporation of various algorithms such as Multinomial Naive Bayes, Decision Tree, and Random Forest. The cross-validation score serves to validate the robustness of the model. Furthermore, the high F1 score of 0.99 under-scores its effectiveness in providing accurate predictions for different classes of diseases. The integration of a user-friendly Flutter mobile application streamlines the procedure of entering symptoms and obtaining disease predictions, enhancing the user experience. In addition to disease prediction, our system enhances its utility by providing supplementary materials from external platforms such as YouTube video suggestions, doctor's blogs, and podcasts through third-party APIs, delivering a holistic understanding of predicted diseases. This research makes notable advancements in bridging the gap between individuals and crucial health information. Our system exhibits potential as a tool for early disease detection, improving health awareness, and contributing to public health and well-being.