a faster boosting-based model made by microsoft I. Basic introduction
LightGBM is a gradient boosting framework that uses a learning algorithm based decision tree.
It can be said to be distributed and efficient, it has the following advantages:
-
faster training efficiency
-
Low memory usage
-
better accuracy
-
Support for parallel learning
-
Can handle large-scale data
Compared with commonly used machine learning algorithms: Speed up
II. LightGBM optimization
Decision Tree Algorithm Based on Histogram
Leaf-wise leaf growth strategy with depth limitation
Histogram acceleration
Direct support for category features (Categorical Feature)
Cache hit rate optimization
Histogram Based Sparse Feature Optimization
Multi-threaded optimization