Breast Cancer Classification - SVM

• Tech Stack: Pandas, Numpy, Python, SVM, Machine Learning, Scikit-Learn, Keras
• Github URL: Project Link

Project Overview: This project focuses on classifying breast cancer using Support Vector Machines (SVM). The goal is to develop a machine learning model that can accurately classify breast cancer as either benign or malignant based on a given set of features. The project leverages a comprehensive dataset of breast cancer cases, performing data preprocessing, exploratory data analysis, feature selection, and model training and evaluation using SVM.

Key Features and Methodology:

1. Data Collection and Preprocessing
Dataset: The project uses the Breast Cancer Wisconsin (Diagnostic) dataset, which includes features computed from a digitized image of a fine needle aspirate (FNA) of a breast mass.
Preprocessing: Data cleaning steps include handling missing values, normalizing feature scales, and splitting the data into training and testing sets.

2. Exploratory Data Analysis (EDA)
Visualization: Various visualization techniques such as histograms, scatter plots, and correlation matrices are used to understand the distribution of data and the relationships between features.
Statistical Analysis: Summary statistics and data distributions are analyzed to gain insights into the dataset.

3. Feature Selection
Correlation Matrix: A correlation matrix is created to identify highly correlated features.
Dimensionality Reduction: Techniques like Principal Component Analysis (PCA) are employed to reduce the dimensionality of the data while retaining significant variance.

4. Model Building
Support Vector Machine (SVM): An SVM model is trained to classify the breast cancer data. Different kernel functions (linear, polynomial, RBF) are experimented with to determine the best-performing model.
Hyperparameter Tuning: Grid search and cross-validation techniques are used to tune hyperparameters and optimize the model performance.

5. Model Evaluation
Performance Metrics: The model's performance is evaluated using metrics such as accuracy, precision, recall, F1-score, and confusion matrix.
ROC Curve: Receiver Operating Characteristic (ROC) curve and Area Under the Curve (AUC) are plotted to assess the classifier's performance.