LightGBM is a gradient boosting framework that using tree based learning algorithms. It can outperform existing boosting tools on both learning efficiency and accuracy. Our [experiments](https://github.com/Microsoft/LightGBM/wiki/Experiments#comparison-experiment) shows that the result of efficiency and accuracy are better than other boosting tools.
LightGBM is a gradient boosting framework that using histogram based tree learning algorithm. It can outperform existing boosting tools on both learning speed and accuracy. Our [experiments](https://github.com/Microsoft/LightGBM/wiki/Experiments#comparison-experiment) shows it is better both on speed and accuracy than other boosting tools.
LightGBM can be run on multiple machines, Our [experiments](https://github.com/Microsoft/LightGBM/wiki/Experiments#parallel-experiment) shows it can perform linear speed up in parallel learning.
LightGBM can leveraging multiple machines to speed-up the training procedure, which can achive linear speed-up in our [experiments](https://github.com/Microsoft/LightGBM/wiki/Experiments#parallel-experiment) settings.
