Research On The Performance Of Steel Industry Listed Companies

Because of the return and the risk of the stock investiment are synchronize,the return is higer with the risk is higer,too.So, it's hard to most of the investors to obtain the maximum return within the limited risk. The numerous individual investors are scanty of needed professional knowledge compare with the minority professional organic investors.Additional,with the lack of time and other objective factors,people can't pay their full attention on the development of stock market instantly. Therefore,they will lost a great deal of opportunities possibly. Consequently, investors are eager to acquire an simple and effective prediction methord which can provide advice when they make their investiment decision. Based on the analysis of operation performance of the listed company of the steel industry,the paper will set up an performance prediction model of steel industry so as to provide useful opinions for investors.The paper selects 36(32 actual) listed companies of the steel industry four years financial affairs datd,then refer to the previous achievement of others and according to the principle of form forecasting indicators system, select 19 financial affairs indicators to compose the forecasting indicators system from the report of listed companies. Next, according to the rough set theory there will obtain 12 indicators as the input of the model after the reduction of information system. Of course,the reduction didn't change the property of the information system mentioned above. And then ,the paper uses Matlab7.0 to establish rough set and neural networks performance prediction model .Then,64 samples are selected to training the model to confirm the parameters, and the remain 32 samples as testing samples to test the model's performance. The results shows that the prediction effect of the combination model is good,the prediction veracity is 71.9%, demonstrate that the prediction model is effective as a useful implement for invesiors. Finally, Finanlly after the comparation of the rough set and neural networks performance prediction model's prediction ability and the normal neural networks model's,we will discover that the prediction veracity is improved from 59.38% to 71.9%.This manifests that the rough set and neural networks performance prediction model is more simple and has more prediction precision than normal neural networks model,and reveals the former's superior adequately.The rough set and neural networks performance prediction model which was established in the paper reveals fine results in the prediction of steel industry listed firms's operation performance,of course, the train of thought in the paper also can provide reference to other industries and other econimics predictions.