Breast Cancer Database Project

This project applies advanced data science and machine learning techniques to the Wisconsin Breast Cancer Dataset (569 records, 33 features). The goal is to build a reliable classification system that helps detect breast cancer at early stages by predicting tumor malignancy with high accuracy. What was done Data Loading & Inspection Loaded 569-row dataset with 33 columns and performed initial shape inspection. Data Cleaning & Outlier Detection Applied Z-Score and IQR methods to detect and clip outliers, achieving 0 remaining outliers after cleaning. Exploratory Data Analysis Generated correlation heatmaps, boxplots, histograms, and class distribution charts to understand the data deeply. Model Training & Evaluation Trained 4 classifiers: Logistic Regression, Random Forest, Decision Tree, and KNN. Evaluated using accuracy scores and confusion matrices.