Prediction On Stock Price Of Agro-related Company

Agriculture is the most important industry in China and it’s the basis of national econ-omy. For the reason that the agro-related companies represents the advanced productivity, the way to keep the stable and progress of the agro-related companies is a essential factor for promoting agricultural development. However, the agro-related companies’development are supported by stock market, therefore, it’s significant in both theory and application for man-agers and invertors to evaluate the value of agro-related companies and predict it’s stock price accurately.There are huge numbers of influence factors of stock price and it’s hard to accurately predict because of the changeable fluctuation. In order to improve the prediction accuracy of agro-related companies’ stock price, this essay carries out study in several aspects:(1) The influence study of agro-related companies’stock price based on agricultural policies. This essay is started with the analysis of the stock price that affected by national policies and agro-related companies’stock price that affected by agricultural policies on a macro level. Then, the occurrence time of the major agricultural policy events were selected as study phase as well as taken Jinjian Cereals Industry Co., Ltd and Long Ping High-Tech etc. as example to analyze the diversity of the increase of the stock price compared to the forward and backward of the occurrence time. The results indicated that the agro-related companies’stock price did response significantly to national’s benefiting-farming policies.(2) One-dimensional time series prediction of agro-related companies’stock price based on ARIMA. In order to predict future tend, One-dimensional time series prediction analyzes the rule between observed values in every period based on historical observations. Time Se-ries and regression analysis of ARIMA model wrere merged into this essay to predict one-dimensional time series data of four agro-related companies’stock price from Aug 11.2014 to Sept 22,2014 including Dahu Aquaculture Co.,Ltd, Jinjian Cereals Industry Co., Ltd,:Hunan Huasheng CO., Ltd and Hunan New Wellful Co., Ltd. The results showed That the average relative error of 4 stock price were less than 3% by utilizing ARIMA and it had certain significance on prediction of stock trend.(3) Multi-dimensional time series prediction of agro-related companies’stock price based on SFS-BPNN. In this easy, the influence factors of stock and other external factors are considered to predict the stock price in practical study. We selected eight factors that consisted of stock’s daily opening, high, low, closing, turnover, volume of transaction as well as the close of daily’s mainland agricultural index and forestry index to forecast tomorrow’s closing price(Y) through empirical research and analysis. Considered the exited information redundancy of eight factors, a new SFS-BPNN model that used to predict stock price was proposed in this easy. The specific procedure was stated below: considered the correlation coefficient value between and tomorrow’s closing price firstly; then, sort factors was intro-duced based on Sequential Forward Search and removed redundant factors to keep the retar-dation factors according to BPNN prediction accuracy of the last five samples of training set; finally, implemented a independent prediction by utilizing the retardation factors. The results of four agro-related companies’ stock price revealed that for the prediction accuracy, SFS-BPNN was better than BPNN and ARIMA and it had high efficiency in predicting agro-related companies’stock price, also, SFS-BPNN can provide effective reference of price prediction for other economic fields.