ABSTRACT
In Chapter 4 (see also Qian & Hua 2004), we presented an analytic frame-work to evaluate individual alpha factors based on the risk-adjusted information coecient (IC). e ratio of average IC to the standard deviation of IC serves as a proxy for the information ratio (IR) of active strategies that employ the alpha factors. We then devoted the next two chapters to the examination of several alpha factors on an individual basis. In practice, alpha models almost always employ multiple factors instead of a single one. So then, the question naturally arises: how to blend these factors optimally into a composite alpha model? e combination of these factors is not restricted to quantitative factors. For instance, some investment rms conduct both fundamental and quantitative researches. How to combine them into a single forecasting process, in terms of ranking or scores, presents a similar challenge.
