The Application Of Clustering In Enterprise Performance Evaluation Based On Period Gene

With the rapid development of our society and market economy, the competition between enterprises is more and more fierce. So whether we can evaluate the enterprise performance timely and reasonably, predict the financial risk on time becomes an imperative problem while managing the micro enterprises and monitoring the macro market. In the era of big data, data mining has the function of mining potential information from large amount of historical data, and will be helpful to solve this problem. This article put forward a new method. The model defined the enterprise period gene firstly, and combined the data discretization and clustering to evaluate the enterprise performance of enterprises.Firstly, the background and meaning of the research was introduced as well as the foreign and domestic research status. The article also analyzed the achievement as well as limitation of them and elaborated the innovative thinking. The second and third part was about the definition and theory of enterprise performance evaluation, financial risk early-warning, data mining, clustering and data discretization. After that, the article introduced the model in detail. The definition of model parameter, selection of sample and index, model building method and model assessment were all described respectively. At last, it made an empirical study, selecting more than 200 companies from 2006-2011 as samples and also 15 indexes as the object of study. Upon examination, the cluster effect was obvious. The rate of accuracy was high. Besides, the article also improved this empirical examination by increasing the weights of key index and limiting the industry. The effect was significant and the rate of accuracy was as high as 87.5%. The model does well in practice, and has the obvious effect in the predicting the risk. As the result, the model is valuable to be spread.