Index Tracking Based on LARS-Lasso and Its Application to Stock Index Future Arbitrage

We use the method of choose covariates of multi-regression to do research on index tracking and stock index future arbitrage,we use LARS algorithm to solve the positive constraint Lasso problem, and give a new method to construct portfolio.And find that:the spot portfolio which used Lasso method has smaller tracking error while using small numbers of stocks.Also,we use the tracking error to study the spot-future arbitrage of HS300,and get good results.