---

🧾 Liver Cirrhosis Stage Detection – Predicting Disease Progression from Clinical Data

Using Random Forest classification to predict liver cirrhosis stages based on patient data from a Mayo Clinic study (1974–1984).

---

📌 Table of Contents

• Overview
• Project Problem
• Dataset
• Tools & Technologies
• Project Structure
• Data Cleaning & Preparation
• Exploratory Data Analysis (EDA)
• Modeling & Evaluation
• How to Run This Project
• Author & Contact

---

Overview

This project predicts the histologic stage of liver cirrhosis using clinical data collected from patients over a 10-year period. The pipeline includes data cleaning, encoding, normalization, exploratory analysis, and Random Forest classification — all built and executed in a Kaggle kernel.

---

Project Problem

Early detection of liver cirrhosis progression can improve treatment outcomes. This project aims to:

• Predict cirrhosis stage (1, 2, or 3) from patient data
• Identify key clinical indicators of disease severity
• Support medical decision-making with interpretable ML models

---

Dataset

• Source: Mayo Clinic study on primary biliary cirrhosis (1974–1984)

|    |   N_Days | Status   | Drug    |   Age | Sex   | Ascites   | Hepatomegaly   | Spiders   | Edema   |   Bilirubin |   Cholesterol |   Albumin |   Copper |   Alk_Phos |   SGOT |   Tryglicerides |   Platelets |   Prothrombin |   Stage |
|---:|---------:|:---------|:--------|------:|:------|:----------|:---------------|:----------|:--------|------------:|--------------:|----------:|---------:|-----------:|-------:|----------------:|------------:|--------------:|--------:|
|  0 |     2221 | C        | Placebo | 18499 | F     | N         | Y              | N         | N       |         0.5 |           149 |      4.04 |      227 |        598 |  52.7  |              57 |         256 |           9.9 |       1 |
|  1 |     1230 | C        | Placebo | 19724 | M     | Y         | N              | Y         | N       |         0.5 |           219 |      3.93 |       22 |        663 |  45    |              75 |         220 |          10.8 |       2 |
|  2 |     4184 | C        | Placebo | 11839 | F     | N         | N              | N         | N       |         0.5 |           320 |      3.54 |       51 |       1243 | 122.45 |              80 |         225 |          10   |       2 |
|  3 |     2090 | D        | Placebo | 16467 | F     | N         | N              | N         | N       |         0.7 |           255 |      3.74 |       23 |       1024 |  77.5  |              58 |         151 |          10.2 |       2 |
|  4 |     2105 | D        | Placebo | 21699 | F     | N         | Y              | N         | N       |         1.9 |           486 |      3.54 |       74 |       1052 | 108.5  |             109 |         151 |          11.5 |       1 |

---

Tools & Technologies

• Python (Pandas, NumPy, Scikit-learn, Seaborn, Matplotlib)
• Kaggle Kernels (Notebook execution and visualization)
• GitHub (Version control and portfolio hosting)

---

Project Structure

liver_cirrhosis_stage_detection/
│
├── README.md
├── liver_cirrhosis.csv                  # Dataset
├── liver_cirrhosis_stage_detection.ipynb  # Kaggle notebook
├── visuals/                             # Plots and charts
│   ├── stage_distribution.png
│   ├── heatmap.png
│   └── confusion_matrix.png

---

Data Cleaning & Preparation

• Encoded categorical features (Sex, Drug, Edema, etc.)
• Normalized numerical features using StandardScaler
• Verified absence of missing values
• Split dataset into training and test sets

---

Exploratory Data Analysis (EDA)

Stage Distribution:

• Balanced across stages 1, 2, and 3

Feature Correlations:

• Strong correlation between Bilirubin, Albumin, and Stage

Visuals:

• Correlation heatmap
• Boxplots for key features
• Stage distribution bar chart

---

Modeling & Evaluation

• Model: Random Forest Classifier
• Accuracy: ~85% on test set
• Evaluation Metrics:
  • Precision, Recall, F1-score
  • Confusion Matrix
  • Feature Importance

---

How to Run This Project

1. Clone the repository:

git clone https://github.com/SBanditaDas/Liver-Cirrhosis-Stage-Detection.git

• Open the notebook in Kaggle or Jupyter: 🔗 liver_cirrhosis_stage_detection.ipynb

• Run all cells to reproduce results and visuals

---

Author & Contact

Sushree Bandita Das

📧 Email: sushreebanditadas01@gmail.com

---