A Fuzzy Logic and Machine Learning Based Approach for Prediction of Skin Cancer

Abstract

Skin disease is among the most regular malignant growth, representing around 5 newlinemillion cases analyzed yearly. Unusual cell improvement on the skin causes skin newlineinjuries, and manual review of skin sores is a troublesome, testing, instinctual, newlineand dreary errand. Because of the diverse and complex properties of skin sore newlinepictures, assessment of skin injury images presents huge hardships. Numerous newlineSC expectation techniques advanced over the most recent twenty years, yet these newlinestrategies were centered basically around ML systems for the forecast. Therefore, newlinein this thesis, we have used the concept of fuzzy logic systems, hybrid systems, newlineevolutionary algorithms, and ML to predict the existence of SC. We used the newlinejoined benefits of the FLS and the ML methods to foresee SC. To satisfy the newlineundertaking, we utilized evolutionary meta-heuristics algorithms for feature selection (FS), automatic fuzzy if-then rules generation algorithm, and fuzzy rules newlinereduction algorithm, which are introduced in the different sections of this thesis. newlineExecution of these algorithms on various datasets shows that the exactness level newlineat each halfway stage is expanded with a general improvement in the accuracy newlineand precision level of SC forecast newline