🩺 Liver Disease Prediction Using Machine Learning

This project uses clinical patient data to predict the likelihood of liver disease using SVM and Naive Bayes models. It is part of a 15-project ML learning sprint aimed at developing healthcare-focused predictive models.

Project Highlights

• Clinical data preprocessing & scaling
• Support Vector Machine (SVM) and Naive Bayes models
• Accuracy comparison with confusion matrices
• Medical interpretation of model outcomes
• Recommendations for future improvements

Dataset

• Source: [Liver Disease UCI Repository or Kaggle version]
• Features include age, bilirubin levels, enzymes, proteins, and diagnosis

Results

Model	Accuracy
SVM	71.55%
Naive Bayes	61.21%

SVM outperformed Naive Bayes, likely due to its robustness in handling correlated clinical features.

Medical Insight

The model identifies Total Bilirubin, Alkaline Phosphotase, and Albumin as critical markers in liver disease prediction. Such models can assist in early diagnosis or decision support in clinical settings.

Files

• liver-disease-prediction.ipynb – Jupyter notebook source code
• liver_patient_dataset.csv – Dataset used
• requirements.txt – Required libraries
• liver-disease-report.pdf – (optional) Final formatted report

Future Improvements

• Use SMOTE for imbalance
• Tune SVM with RBF or polynomial kernels
• Try ensemble models (XGBoost, Gradient Boosting)

Author

Developed by Azib Malick
© 2025 Azib Malick. All rights reserved.