The statistical estimation of the mean and variance of the asset returns are particularly important for portfolio selection. When the dimension of the variables are high, the conventional estimation methodologies are likely to break down. In this project, we explore various statistical methodologies in an attempt to obtain estimation with improved accuracy.
The applicant will read relevant articles published recently in statistics journals and present some of them. He/she will carry out a large scale numerical work, including simulation and collection and analysis of real data.
The applicant will learn some new statistical techniques.