Excavating Excess Returns Of Equity Incentive Events Based On Random Forest Regression Model

There are numerous investment strategies in the secondary market,of which the event driven strategy and the factor selection strategy have long been approved by investors.This thesis selects the equity incentive plan which is mature in China as an event.The equity incentive plan links the future earnings of the company's core personnel with the company's future performance to encourage employees.In theory,by applying the equity incentive plan,the corporate performance will grow.Therefore,when the listed company releases the equity incentive plan,the market will react immediately and stock price fluctuates.According to the statistics by Wind,there is indeed an excess yield relative to the CSI 500,and the effect of equity incentive on the preplan day is stronger.The effect of equity incentive was most pronounced in 2015 and almost disappeared in 2016 and 2017.Therefore,based on the event-driven strategy and the factor stock selection,this thesis constructs an event-driven multi-factor selection model to excavates excess returns from equity incentive events.From the experience of investment,the internal factors(incentive target,incentive limit and listed segment)and external factors(valuation factor,profitability factor,operation factor and dividend factor)of equity incentive time are selected.The data between 2010 and 2016 was used as a training set and data of 2017 as a test set,and a random forest model constructed by Python was brought in.In the end,the average accuracy of the prediction is 55.92%and based on the results obtained from the test set.The rate of average return on the strategy is 200.56%and the rate of excess return relative to the CSI 500 is 95.74%,which far exceeds the general strategy and model.