Human disease prediction and doctor recommendation system using machine learning approaches: A case study in Bangladesh

Abstract

Despite significant strides in modern technology, a substantial portion of the global population is still not getting proper medical care. This issue is particularly pronounced in developing countries which face a double burden of communicable and non-communicable diseases. On the other hand, limited access to quality healthcare, particularly in remote areas where skilled doctors may not be available, is a significant challenge for patients to receive appropriate treatment. To solve this problem, this project will be able to make the treatment much easier and more accurate. The system is designed to enhance the accuracy, reliability, and efficiency of disease prediction by leveraging the power of machine learning algorithms and reliable datasets. It will also make the work of doctors a lot easier because this project can diagnose possible diseases and give suggestions about what restrictions should be followed at home to get rid of these diseases. Since disease prediction is a very crucialsubject where accuracy has to be maximized, ten machine-learning models, including Decision Trees, Random forest, Bagging Classifier, Support Vector Machine and AdaBoostClassifier and hybrid (a combination of several machine-learning) models have been used in this project that capitalize on the strengths of models or data types while mitigating their respective weaknesses. The fundamental objective of this work is to significantly enhance the user experience of the existing disease prediction system. This is achieved through the design and implementation of a user-friendly interface using a Python framework "Streamlit" that facilitates not only disease prediction but also recommends appropriate doctors and suggests preventative measures