Research On Quantitative Stock Selection Model In China Market

Quantitative investment is an effective method that has been proved by much investment practice in recent years. In this pape, we build A-share stock selection models, based on the momentum and reversal effect of stock market in China, combined with Alpha’s stock picking strategy and clustering algorithm. This paper also undertakes a back-test that based on the stock data from 2011-1-18 to 2015-2-17.Firstly, build a α momentum stock picking strategy- Model I combined with the actual situation of China’s stock market. In the meantime, we define the stock selection factor μ.Then the clustering algorithm was used to construct China stock market’s quantitative stock selection Model II. However, through the simulation data we find the the relatively lower operation efficiency greatly limit the number of clustering stock. So, we use the k- center of rotation point in this paper which the complexity is O(n2k2). Finally, we build Alpha clustering momentum model by combining the above two kinds of model. We verify the validity of the model by comparing the relative to the CSI 800 index excess return and the positive and negative abnormal returns appear frequency, cumulative earnings results.By the empirical test, we can see that the Model I doesn.t get the excess returns and super large frequency relatively to the CSI 800 index. But in the bull market we basically have 40 % of the excess proceeds under different(u, p) parameters and we obtained by over176.04% annual yield under the optimal parameter combination. Model II exceeded the yield of the CSI 800 index and the pure momentum reversal model firmly. Through the comparison of the clustering- Alpha momentum model, Model I and Model II, we can draw a conclusion that the clustering- Alpha momentum model has a stronger ability to resist risks and improves the frequency and amplitude of positive excess returns. The empirical results demonstrate the effectiveness of the selection model in China stock market.