Generalization of change-point detection in time series data based on direct density ratio estimation

The goal of the change-point detection is to discover changes time series distribution. One state art approaches based on direct density ratio estimation. In this work, we show how existing algorithms can be generalized using various binary classification and regression models. particular, that Gradient Boosting over Decision Trees Neural Networks used for purpose. are tested several synthetic real-world datasets. results proposed methods outperform classical RuLSIF algorithm. Discussion cases where have advantages also provided.