Comparative Study of Fuzzy Classification Methods on Breast Cancer Data

Title: Comparative Study of Fuzzy Classification Methods on Breast Cancer Data

Abstract: This research study investigates the performance of four fuzzy rule generation methods on Wisconsin breast cancer data. The methods generate fuzzy if-then rules using different approaches, such as mean and standard deviation, histogram of attributes values, certainty of each attribute into homogeneous fuzzy sets, and overlapping areas partitioning. The study aims to determine the most effective approach for classifying breast cancer data. Simulation results show that the Modified grid approach has a high classification rate of 99.73%, making it the most effective method for fuzzy classification of breast cancer data.

Keywords: Fuzzy systems, Breast cancer diagnosis, Rule generation methods, Fuzzy if-then rules, Homogeneous fuzzy sets, Overlapping areas partitioning, Breast cancer data classification, Fuzzy logic applications in medicine

Link to Article: https://arxiv.org/abs/0405008v1 Authors: arXiv ID: 0405008v1