HCC Survival Prediction

Survival Prediction for Hepatocellular Carcinoma (HCC)

Project Description

This project was developed as part of the curricular unit "Elements of Artificial Intelligence and Data Science" of the BSc in Artificial Intelligence and Data Science.
It was graded 20/20.

The objective is to develop a complete data science pipeline to predict the one-year survival of patients diagnosed with Hepatocellular Carcinoma (HCC).

Project Structure

The project is divided into the following main stages:

1. Data Exploration

• Exploratory data analysis
• Examination of feature types
• Class distribution
• Attribute-level values
• Identification of data inconsistencies

2. Data Preprocessing

• Missing value imputation
• Data transformation and scaling
• Feature engineering

3. Data Modeling (Supervised Learning)

• Selection of classification algorithms
• Definition of training and testing sets
• Model performance evaluation

Algorithms used:

• Decision Trees
• K-Nearest Neighbors (KNN)
• Random Forest
• Gradient Boosting
• Multi-Layer Perceptron (MLP)
• Logistic Regression
• Stacking Classifier
• Support Vector Classifier (SVC)

4. Model Evaluation

• Comparison of classification results using standard metrics:
  • Confusion Matrix
  • AUC/ROC
  • Precision
  • Recall
  • Accuracy

5. Results Interpretation

• Extraction of meaningful insights
• Explanation of model behaviors
• Recommendations for future analyses

Execution Instructions

Prerequisites

Ensure you have Python 3 installed.
All required libraries can be installed from the requirements.txt file.