Implementation of machine learning techniques for disease diagnosis. (2022)

• Record Type: Journal Article
• Authors: Mall, Shachi; Srivastava, Ashutosh; Mazumdar, Bireshwar Dass; Mishra, Manmohan; Bangare, Sunil L.; Deepak, A.
• Is Part Of: Materials today. Volume 51:Part 8(2022)
• Page Start: 2198
• Page End: 2201
• Publication Date: 2022
• Subjects: Data Mining -- Machine Learning -- Classification -- Decision Tree -- Prediction -- Heart Disease; Materials science -- Congresses -- Periodicals; 620.1
• DOI: 10.1016/j.matpr.2021.11.274
• Languages: English
• ISSNs: 2214-7853

Abstract

Abstract: Recently, data mining and machine learning techniques have found widespread use in the field of healthcare. The objective of this study is to develop an automated method for diagnosing illnesses. A Fuzzy logic-based random forest approach and a thorough examination of the patient's medical records are used to diagnose the disease. Clinical diagnosis is performed with the aid of a doctor's expertise and understanding in traditional healthcare. It is more challenging to provide good healthcare in rural and remote areas because patients are more likely to travel a long distance to visit a specialist. Because the number of medical practitioners and facilities in these areas is limited, providing an expert diagnosis in a fair period of time is challenging. The problem can be solved by using expert systems for disease diagnosis that employ data mining techniques and fuzzy logic. Decision trees are often used in machine learning to predict outcomes. Fuzzy datasets are an excellent choice for describing medical facts and expert opinions. Fuzzy decision trees build simple decision trees using fuzzy input. In this proposed system, an expert system that diagnoses disease using a random forest algorithm and fuzzy decision trees is provided. The fuzzy decision trees increase the accuracy of the diagnostic system. On the UCI repository, the proposed method is assessed and found to be more efficient in sickness prediction than current strategies. Classification accuracy has risen as temporal complexity has decreased.