Multi-factoc Stock Selection Model Based On The Combination Of Quantity And Price Data

In recent years,quantitative investment has become one of the mainstream investment methods at home and abroad.At present,most researches on multiple factors focus on fundamentals or simple volume and price factors.However,with the improvement of market effectiveness,these factors are gradually losing their effectiveness.Therefore,it is of great practical significance and value to study the stock selection effect of complex technical factors represented by the combination of quantity and price characteristics.Based on the 101 combination factors of quantity and price characteristics published by WorldQuant.This paper selects the factors that are effective in domestic stock market.Then we construct the strategy of multi-factor stock selection,and conduct a back test.This paper tested the effectiveness and the prediction cycle of all the 101 factors in domestic stock market.Then we select 16 effective factors and combine these effective factors linearly.We predict the return rate of individual stocks in the next two days by the model.The results show that the model has good prediction ability.Under the condition of daily frequency,the model meets the standard of excellent multi-factor model.This paper discusses three control methods of turnover rate.And we choose the method which the control of turnover is at the level of objective function.Then we construct the optimization problem under the constraints of style and industry neutrality.And we use the least square programming method to solve the problem.The historical back-test's results show that a larger market fluctuation will give the strategy larger alpha space,but the overall style of strategy combination is independent,which can produce considerable returns under different market styles.Brinson attribution shows that the strategy return mainly comes from stock selection effect.The calculation strategy can contain more than ten million yuan.