Accelerated-Gradient-Boosting
Gradient Boosting model are suppose to provide stellar performance for supervised learning.Recently it has been proven that we can have faster convergence with Nestrov's momentum sequence for optimization.This is the paper.
Author has shown the results in R. This is the python implementation of the same.
Following is the algorithm:-
We have included two datasets(https://archive.ics.uci.edu/ml/datasets.html)
- Crime data
- CBM data
Results
Crime Data Non Acceleration Performance
Crime Data Acceleration Performance
CBM Data Non Acceleration Performance
CBM Data Acceleration Performance
