An accurate soft diagnosis method of breast cancer using the operative fusion of derived features and classification approaches
Sunil Kumar Jha, Jinwei Wang, Raju Shanmugam · Expert Systems · 2022
Abstract Breast cancer is one of the most common types of cancer around the world. The early‐stage recognition of breast cancer is favourable for the diagnosis and treatment of the affecting patient. Data mining approaches can ease the diagnosis of breast cancer by analysis of the associated dataset for real‐time decision‐making. The present study proposes an effective transformation approach of experimental attributes of a breast cancer dataset using the latent semantic analysis and the fusion of derived features with classification methods for accurate recognition of breast cancer. The proposed approach is validated using the most widely used benchmark and open‐access breast cancer dataset. The transformed features of the original dataset result in 100% recognition efficiency using a multilayer perceptron, support vector machine, multi‐class classifier and functional tree classifiers. Other classifiers, like naïve Bayes, rotation forest, simple linear logistic regression and logistic model tree result in recognition accuracy between 96.85% and 99.30% using a similar feature subset. Besides, the optimal subset of derived features has been affirmed based on the evaluation metrics of classification approaches.