A Behavioral Classification of Investors in Shanghai Stock Market

In our study, we analyzed a securities clearing dataset from a Chinese brokerage firm, and classified investors based on their behavior. We calculate every account´s market value, number of stocks and trading frequency, then reduce every account´s stock portfolio to corresponding expected risk and rate of return. And we cluster these accounts using k-means algorithm, and obtain four distinct clusters, which are small-account investors, big- account investors, medium-account with risk-aversion investors, medium-account with risk-preference investors. The analysis of the results reveals that small account investors prefer risks a little more than that with scale capital. Furthermore, among the investors with the medium-account, more people tend to avert risk, and risk- preference investors are more inclined to change their holding stocks than risk-aversion ones.