AI Disease Prediction System

An AI-powered disease prediction system that implements predictive diagnosis models for multiple diseases using data munging, feature selection, model training, and accuracy calculation. This project validates and improves existing methods in the disease diagnosis pipeline.

---

Table of Contents

• Overview
• Features
• Project Structure
• Diseases Covered
• Methodology
• Technologies Used
• Installation & Setup
• Usage
• Datasets
• Results
• Contributors
• License

---

Overview

This project implements an AI-based predictive diagnosis system focusing on improving key aspects of the disease prediction pipeline:

1. Data Munging - Replacing missing values using k-nearest neighbors (KNN) instead of mean substitution to reduce bias
2. Feature Selection - Using simulated annealing algorithm with adaptive p-value thresholds
3. Model Training - Training predictive models on medical datasets
4. Accuracy Evaluation - Comprehensive cross-validation and accuracy calculation

---

Features

• Multi-disease prediction support (Breast Cancer, Heart Disease, Diabetes)
• KNN-based data imputation for handling missing values
• Simulated annealing for optimized feature selection
• Adaptive p-value significance thresholds
• Data augmentation techniques (noise addition, scaling)
• Cross-validation support
• MySQL database integration
• Comprehensive result visualization with graphs

---

Project Structure

AI-Disease-Prediction-system/
├── BreastCancer/                 # Breast cancer dataset files
├── BreastCancerPrediction/       # Breast cancer prediction module (Maven)
├── Diabetes/                     # Diabetes dataset files
├── HeartDiseaseData/             # Heart disease dataset files
├── HeartDiseasePrediction/       # Heart disease prediction module (Maven)
├── Presentation/                 # Project presentation files
├── Project Report/               # Detailed project report
├── Results - Graphs/             # Result visualizations and graphs
├── commons-math3-*.jar           # Apache Commons Math library
├── mysql-connector-java-*.jar    # MySQL JDBC driver
├── build.xml                     # Ant build configuration
├── manifest.mf                   # JAR manifest file
└── README.md                     # Project documentation

---

Diseases Covered

Disease	Dataset Source	Prediction Module
Breast Cancer	Wisconsin Breast Cancer Dataset	BreastCancerPrediction.java
Heart Disease	Cleveland & Hungarian Heart Disease Datasets	HeartDiseasePrediction.java
Diabetes	Pima Indians Diabetes Dataset	Diabetes/ folder

---

Methodology

Data Munging Improvements

• Replaced mean imputation with k-nearest neighbors (KNN) method
• Reduced bias introduced by majority-class substitution

Feature Selection

• Implemented simulated annealing algorithm for optimal feature selection
• Adaptive p-value thresholds within configurable upper and lower bounds
• Configurable initial temperature and cooling rate parameters

Model Evaluation

• 90/10 train-test split method
• Cross-validation results provided throughout the report
• Accuracy comparison across different feature sets

Data Augmentation

• Noise addition techniques
• Data scaling methods
• Improved dataset diversity

---

Technologies Used

• Programming Language: Java
• Build Tools: Maven, Apache Ant
• Database: MySQL
• Libraries:
  • Apache Commons Math 3.x
  • MySQL Connector/J 8.0
• IDE: NetBeans
• Version Control: Git & GitHub

---

Installation & Setup

Prerequisites

• Java Development Kit (JDK) 8 or higher
• MySQL Server (local installation)
• Maven (for prediction modules)
• Apache Ant (for main project)
• NetBeans IDE (recommended)

Steps

1. Clone the repository:

   git clone https://github.com/code-divyu/AI-Disease-Prediction-system.git
   cd AI-Disease-Prediction-system

2. Configure MySQL Database:

   • Open DBConnection.java in both BreastCancerPrediction/ and HeartDiseasePrediction/ folders
   • Update the MySQL username and password:

     String username = "root";  // Change to your MySQL username
     String password = "";      // Change to your MySQL password

3. Build the project:

   • For Maven modules:

     cd BreastCancerPrediction
     mvn clean install

   • For the main project:

     ant build

---

Usage

Running the Prediction Modules

1. Breast Cancer Prediction:

   cd BreastCancerPrediction
   java -cp target/classes com.mycompany.breastcancer.BreastCancerPrediction

2. Heart Disease Prediction:

   cd HeartDiseasePrediction
   java -cp target/classes com.mycompany.heartdisease.HeartDiseasePrediction

Note

The main functions in each program do not execute all functionalities in sequence by default. Uncomment specific function calls to run individual modules as needed.

---

Datasets

Breast Cancer Dataset

• Source: UCI Machine Learning Repository - Wisconsin Breast Cancer Dataset
• Location: BreastCancer/ folder
• Files: breast-cancer-wisconsin.data, wdcb.data, Index

Heart Disease Dataset

• Source: UCI Machine Learning Repository - Heart Disease Datasets
• Location: HeartDiseaseData/ folder
• Files: processed.cleveland.data, hungarian.data, heart-disease.names

Diabetes Dataset

• Source: UCI Machine Learning Repository - Pima Indians Diabetes Dataset
• Location: Diabetes/ folder

---

Results

Result visualizations and graphs are available in the Results - Graphs/ folder. The project report contains detailed accuracy metrics, cross-validation results, and comparative analysis of different methods.

---

Contributors

• Divyanshu (@code-divyu)

---

License

This project is open-source and available under the MIT License.

---

Note: This project is intended for educational and research purposes. It should not be used as a substitute for professional medical diagnosis.