A Hierarchical Clustering Random Forest Model Approach to Diagnose the Breast Cancer Diseases

Breast cancer poses a grave danger to women's health, accounting for the biggest number of cancer-related fatalities among females. Accurate identification and successful treatment are essential for reducing the breast cancer mortality rate. Widespread use of machine learning, namely Random Forest (RF), for accurate illness detection. Nevertheless, poor classification performance and substantial similarity across training decision trees may hinder the classification performance of the RF model. In a Hierarchical Clustering Random Forest (HCRF) model, sample trees with low similarity and high precision were selected using hierarchical clustering. Variable Importance Measure (VIM) enhanced breast cancer predictive characteristics using two datasets from the University of California Irvine (UCI) Machine Learning repository, the proposed method was evaluated for accuracy with 97.05 and 97.76 percent in HCRF with VIM feature selection on both datasets. The suggested approach is an efficient diagnostic tool for breast cancer.