Skip to content
/ mlrap Public
forked from NianSan-H/mlrap

Machine Learning Regression Analyse Packages

License

Notifications You must be signed in to change notification settings

obaica/mlrap

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

30 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

mlrap: Machine Learning Regression Analyse Packages

Table of Contents

Introduction

MLRAP (Machine Learning Regression Analysis Package) is designed to assist materials scientists in constructing regression models and establishing structure-property relationships. This software offers a user-friendly command-line interface and automates the machine learning workflow in Python. It encompasses modules for data preprocessing, feature engineering, model evaluation, model optimization, and model interpretation. This automation significantly reduces the workload for users and enhances the efficiency of materials research and development.
fig

Installation

MLRAP supports Python 3.8+. It is recommended to install it within a virtual environment using virtual environment management tools.
First create a conda environment: Install miniconda environment from https://conda.io/miniconda.html based on your system requirements. Then, create a virtual environment:

conda create -n mlrap python==3.8.0
conda activate mlrap

Extract the compressed archive to a directory, and then execute the following command:

python setup.py install

We highly recommend installing in editable mode:

pip install --editable .

Usage

After installation, execute the command mlrap -h to view the help interface.

Usage: mlrap [OPTIONS] COMMAND [ARGS]...

  Machine learning regression analyse packages

Options:
  -h, --help  Show this message and exit.

Commands:
  run     Global run base config file.
  subrun  Run step by step.

If you have already prepared your dataset (a CSV file containing chemical formulas and target properties), simply execute the command mlrap run train in the directory where your dataset is located. MLRAP will automatically perform feature engineering, hyperparameter optimization, and model training for you. Process data will be output in CSV format, and five images will be generated as follows:

fig

Note

During hyperparameter optimization, we employ Bayesian hyperparameter optimization, where the best hyperparameters are selected for the final model training.

Contributors

About

Machine Learning Regression Analyse Packages

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%