Healthcare-Analytics-Toolset

A comprehensive machine learning project focused on breast cancer detection and healthcare cost prediction. Features multiple models including CNN image classification for FNA samples, decision tree analysis for patient data, and linear regression for insurance cost forecasting. Built with TensorFlow, scikit-learn, and pandas.

Project1

This component implements a linear regression model to predict healthcare insurance costs based on various patient factors. Features:

Processes demographic and health-related data Converts categorical variables (sex, smoking status, region) to numerical values Implements linear regression for cost prediction Provides detailed feature importance analysis Visualizes predictions vs actual costs Generates coefficient analysis for factor impact

Dependencies:

pandas numpy scikit-learn matplotlib

Project2

A decision tree-based classification system for breast cancer diagnosis prediction using patient data. Features:

Handles multiple categorical variables (Race, Marital Status, Cancer Stages) Implements decision tree classifier with depth optimization Provides performance metrics (accuracy, log loss) Generates visual decision tree representation Includes model performance visualization Cross-validates results

Dependencies:

pandas numpy scikit-learn matplotlib

Project3

A Convolutional Neural Network (CNN) based system for analyzing Fine Needle Aspiration (FNA) images for breast cancer detection. Features:

Processes and normalizes medical images Implements CNN architecture for image classification Provides real-time training monitoring Generates performance visualizations Includes prediction capability for new images Saves trained models for future use

Dependencies:

TensorFlow OpenCV (cv2) numpy matplotlib

Data Requirements

Insurance data: CSV file with columns for age, sex, BMI, smoking status, region, and charges Cancer diagnosis data: CSV file with patient information and cancer status Image data: Organized in directories with benign and malignant FNA images

Output

Each component generates:

Performance metrics Visualization plots Model evaluation results Prediction outputs