BMI Prediction Using Metagenomic Data

This project focuses on predicting Body Mass Index (BMI) using microbiome data through various linear machine learning regression models.

Project Structure

Overview

The project implements machine learning models to predict BMI values using metagenomic data. It includes data preprocessing, model training, evaluation, and overfitting analysis.

Key Components

• Data Processing: Cleaning and preprocessing of metagenomic data
• Model Training: Implementation of various regression models
• Evaluation: Performance assessment using metrics like R² and RMSE
• Overfitting Analysis: Detection and visualization of model overfitting

Getting Started

Installation

1. Clone the repository: git clone [repository-url] cd Assignment-1

2. Install required packages: micromamba env create -f environment.yml

Usage

1. Data Preprocessing:

   • Located in data/ directory

2. Model Training:

   • Navigate to notebooks/model_analysis.ipynb
   • Runs multiple regression models on the preprocessed data
   • Evaluates model performance on validation set

3. Analysis:

   • Feature importance visualization
   • Model performance comparison
   • Overfitting detection and analysis