Regularization and Feature Selection in Least Squares Temporal Difference Learning
Description
This is Python implementations of Least Angle Regression Temporal Difference (LARS-TD) algorithm and Least-Squares Temporal Difference (LSTD). For more information on the algorithm please refer to the paper
“Regularization and Feature Selection in Least Squares Temporal Difference Learning”
https://zicokolter.com/publications/kolter2009regularization.pdf
In this paper, the authors tried to propose a regularization framework for least-square temporal differences learning. Specifically, they presented an approach to find the fixed point by using l1 regularization framework. To evaluate the framework’s efficiency, they examined the framework by using two well-known problems, which means Mountain Car and Chain Domain. The results showed that the framework could deal with challenges well
Executing program
python main.py
