5. More on Stock Selection Strategy: Alpha Hunting, Risk Adjustment, and Nonparametric Diagnostics

In the previous chapter, we introduced a general procedure and multi-factor framework for alpha construction of stock selection strategies. In this chapter, we continue to explore stock selection strategy with more advanced topics. In particular, we focus on alpha (new factor) hunting, risk adjustment, and nonparametric diagnostics. Regarding new alpha discovery, we present the guidance of IPARE. From a methodological perspective, we introduce the weighted least squares (WLS) method, which provides a tool to integrate risk into a multi-factor alpha model. We then introduce nonparametric approaches as a complement to parametric analysis. In the industry insights section, we provide a nonparametric diagnostics package used in the industry to investigate a new factor. The last section on R programming shows how to refine plots with parameters.