Optimal Investment Strategy Based On Hypothesis Testing

Sparse assets selection and portfolio investment are discussed in this thesis. Based on Antoine’s[7] one-side P-value investment rules, two aspects are expended. Firstly, two-sides P-value investment rule is proposed according to one-side P-value investment rule. That is to say, maximize the possibility of the investor’s goal falling into some certain range [a, b] in order to get the optimal investment strategy. On the one hand, two-sides P-value investment rule inherits advantages of solving estimation risks from one-side P-value investment rule. On the other hand, the simulation result shows:if investors’ goal is low, the Sharpe ratio of two-sides investment rule is better than one-side rule’s. Secondly, Antoine didn’t discuss sparse portfolio selection, but in practice, it’s very important to select some proper assets from a large of assets. According to Antoine’s framework, sparse portfolio selection model based on hypothesis testing is raised. Under premise of forbidding short selling and controlling the lower limit of the possibility of null hypothesis being accepted, minimize the number of investing assets so that realize sparse portfolio selection. The simulation result shows under premise of forbidding short selling, the performance of this model is better than Fan’s[11] investment model.