Exploring Dimensionality Reduction Techniques for Improved Breast Cancer Diagnosis

A crucial area of medical study is the diagnosis of breast cancer, where managing the inherent complexity of high-dimensional information poses a challenge in addition to precise identification. In order to improve diagnostic accuracy, this research investigates dimensionality reduction strategies. This study’s main goal was to improve the accuracy and interpret ability of breast cancer diagnosis by using dimensionality reduction techniques. The goal of the study is to find significant patterns for useful diagnostic models by examining how preprocessing methods affect a high-dimensional dataset. Starting with a dataset including 569 observations and 30 attributes, careful examination reveals imbalances in the dataset (63% benign, 37% malignant). We used Pearson correlation coefficients to detect and eliminate highly correlated features in order to address multi collinearity. A subsequent adjustment of the data using min-max normalization guarantees consistent weighting. Then, for thorough dimensionality reduction, Principal Component Analysis (PCA) is employed. Screep lots and biplots are used to visually represent data, highlighting how well-suited early principle components are for separating benign from malignant instances. Our findings confirm the effectiveness of the procedure by showing a significant 24% decrease in data dimensionality. This work highlights the critical role that dimensionality reduction plays in improving breast cancer diagnosis for more precise, effective, and understandable models, and it calls for further investigation of the specific findings.