Machine Learning

for tabular data

mljar builds a complete Machine Learning Pipeline
+ advanced feature engineering
+ algorithms selection and tuning
+ automatic documentation
+ ML explanations

Machine Learning for Everyone

There are four built-in modes in the mljar AutoML framework.
You can also create your own custom modes.

Explain mode


Perform exploratory analysis, search for a signal in the data, and discover relationships between features in your data with AutoML

Read More »
Compete AutoML mode


Train top ML models with advanced feature engineering, many algorithms, hyper-parameters tuning, Ensembling, and Stacking

Read More »
AutoML Perform mode


Stay a head of competitors and predict the future whith advanced ML. Deploy your models in the cloud or use them locally

Read More »
AutoML Optuna mode


Highly tune ML models (LightGBM, CatBoost, Xgboost, Random Forest, Extra Trees) with Optuna framework and easy MLJAR AutoML API

Read More »

The Most Advanced Feature Engineering

Understand your data

Features Preprocessing

Don't need to worry about data conversions into numeric or missing values. AutoML will handle this for you!

Read More »

Golden Features Search

Search for new features with great prediction power. Combine existing data into new Golden Features.

Read More »

K-Means Features

Ehance your data with new features created with K-Means clustering algorithm.

Read More »

Features Selection

Use only relevant features. Automatically remove noisy features in your ML pipeline.

Read More »

Build Better ML Models Today

Automatic Exploratory Data Analysis

Discover your data properities with automatically created data visualizations.

Read More »

Early Stopping

Stop model training exactly when it is needed and avoid overfitting with early-stopping.

ML Explainability

Understand your data and models with Machine Learning Explanations. Decision Tree plots, Features Importance, SHAP explanations, and more ...

Read More »

Models Ensembling

Achieve great predictive accuracy with advanced ensembling techniques. Stack and combine models for performance improvements.

Read More »

Track Model's Metrics

Track your model's metrics and results. Get full reproducibility for your Machine Learning experiments.

Boost on Errors

Improve your model performance by learning on previous model's errors. Boost sample weight values on difficult data points.

Automatic Documentation

Reports Generation

Results of the ML analysis are available as Markdown or HTML reports. You can check there: hyper-parameters, metrics, learning curves, plots, explanations.

Read More »

Perfect for Notebooks & GitHub

The Markdown and HTML reports work perfect with Jupyter Notebooks, Kaggle Notebooks and GitHub. Present your results with elegance.

Read More »
Automated Documentation in Markdown and HTML

State-of-the-art AutoML Performance

The mljar provides the state-of-the-art performance on binary, multiclass classfication, and regression.
Check the comparison of the popular AutoML frameworks on Kaggle datasets.

Supported Machine Learning Tasks

Binary Classification

Binary Classification

Multi-class Classification

Multi-Class Classification



Easy installation and usage!

You can install MLJAR AutoML Python package
with PyPi repository, from source or run in the docker container.