A Study Upon The Changes Of Earnings Rate After The Issuing Of Increase And Bonus

Increase and bonus have recently become popular ways of dividend issuing in China’s securities market. The influence to the earnings rate after the issuing of increase or bonus is caused by a lot of factors. This paper collects all of the496companies who has issued increase or bonus during the period of June,2012to December,2012and tries to figure out the factors which have brought shocks to the earnings rate of these companies, with the methods of data mining, then tries to explain the modeling results with practical laws of the market. In addition, the shocks mentioned include not only the short-term ones but also the mid-term and the long-term ones.By using the methods of logistic regression, CART, and neural network, this paper has made conclusions on both aspects of methods and modeling results.First, the modeling results show something regular among these companies. What’s more, most of them can be explained by practical laws of the securities market thus the research work achieved by the data mining methods is practical and meaningful.Here follows the useful conclusions extracted from the modeling results:1. Most of the companies who issued bonus stocks are expected to get positive return during all the three time scales.2. There exists a positive relationship between the earnings rates in the half-year report and those among the time scales.3. The market prefers those companies whose price earnings ratio falls into one certain scale. And those with a price earnings ratio that is too high or too low are more probably to get a negative earnings rate.There are two things that have made this paper unique. First it imports the methods of data mining to research the characteristic of the securities market. Secondly it combines the results with the real market in order to find out some reasonable advices and strategies upon real investment operating. Thus the paper not only makes a meaningful trail upon the appliance of data mining methods in the securities market, but also provides some valuable references for the investment in China’s stock market.However there still remains lots of space for this paper to improve. First, more input variables might improve the performance of the models. Second, the paper is insufficient to illustrate the relationship between insider trading and the changes of the earnings rate. Finally there is a lot to do in order to figure out the sudden shock caused by the insider trading.