CharlesHendrix

CharlesHendrix is a project that uses a genetic algorithm to generate musical compositions. The fitness of each composition is evaluated by a pre-trained machine learning model, guiding the evolutionary process toward creating aesthetically pleasing music.

Project Overview

This project combines concepts from evolutionary computation and machine learning to explore algorithmic music generation. The core of the project is a genetic algorithm that evolves populations of musical "genomes." A machine learning model, trained on user feedback, acts as the fitness function, predicting a score for each generated genome.

Directory Structure

The project is organized into the following directories:

• data/: Contains raw and processed data, including user feedback, generated compositions, and feature notes.
• documentation/: Holds project documentation.
• model/: Stores the serialized machine learning model, scaler, and other necessary objects.
• src/: Contains all the Python source code.
  • data_processing/: Scripts for data extraction and feature engineering.
  • explainability/: Tools for interpreting the machine learning model's predictions.
  • genetic_algorithm/: The core implementation of the genetic algorithm, including fitness evaluation and genome handling.
  • test_models/: Scripts for evaluating and comparing different machine learning models.

Getting Started

Prerequisites

• Python 3.10 or higher
• pip

Installation

1. Clone the repository to your local machine:

   git clone https://github.com/SugarStoneMaster/CharlesHendrix.git
   cd CharlesHendrix

2. Install the required Python packages using requirements.txt:

   pip install -r requirements.txt

Usage

To run the genetic algorithm and generate a new musical composition, execute the main script:

python src/main.py

The script will run the evolutionary process and save the resulting composition.

How It Works

1. Genetic Algorithm: The process starts with an initial population of random musical genomes.
2. Fitness Evaluation: Each genome is converted into a set of musical features. These features are fed into a pre-trained machine learning model (model/model.joblib) which outputs a fitness score.
3. Selection, Crossover, and Mutation: Genomes with higher fitness scores are more likely to be selected for reproduction. New genomes are created through crossover (combining parts of two parent genomes) and mutation (randomly altering a genome).
4. Evolution: This cycle repeats for a set number of generations, with the population's overall fitness improving over time. The best genome from the final generation is selected as the result.