A Comparative Study on a Disease Prediction System Using Machine Learning Algorithms

One of the most essential and fundamental factors that motivates people to seek assistance is their physical well-being. A vast range of ailments affect people nowadays, making them extremely vulnerable. Thus, disease prediction at an early stage has now become increasingly relevant as a result of these developments. Machine Learning is a relatively new technique that can aid in the prediction and diagnosis of diseases. To diagnose respiratory difficulties, heart attacks, and liver disorders, this study employs machine learning in conjunction with symptoms. All these diseases are the focus of our investigation because they are extremely prevalent, incredibly expensive to treat, and impact a significant number of people at the same time. A variety of supervised machine learning methods, including Naive Bayes, Decision Trees, Logistic Regression, and Random Forests, are used to forecast the disease based on the provided dataset. The discussion of learning categorization based on correctness ends with this conclusion. Flask is also used to construct a platform that allows visitors to forecast whether they will contract a specific illness and take appropriate precautions if they do contract the illness. Among the most important aspects of healthcare informatics is the prediction of chronic diseases. It is critical to diagnose the condition at the earliest possible opportunity. Using feature extraction and classification methods for classification and prognosis of chronic diseases, this study gives a summary of the current state of the art. The selection of elements that are appropriate for a classification system is critical in improving its accuracy. The decrease of dimensionality aids in the improvement of the overall performance of the machine learning system. The use of classification algorithms on disease datasets gives promising results in the development of adaptive, automated, and smart medical diagnostics for chronic diseases, according to the researchers. Using parallel classification systems, it is possible to speed up the process while also increasing the computing efficiency of the final findings. This paper provides a complete analysis of several feature selection strategies, as well as the advantages and disadvantages of each method.