The Application Of Data Mining Technology In The Stock Forecast

Stock market plays an essential role in securities and financial industry. It also attracts increasing attention from the investors. Valid stock forecast is of great importance in financial investment field. Hence, making analysis and forecast on stock prices has extraordinary theoretic significance and practical value. However, due to the various complicated factors influenced by policy, economy and investors' mentality, there is no denying the fact that those uncertain factors bring great difficulty to the forecast of the stock.Up to the present moment, data mining technology has already shown its great vitality of further development. Some people even compare the influence of this technology with the invention of internet and fire. Its findings have also been applied into many industries such as finance, medical care, retailing, manufacturing, engineer and science industry etc. While the particularity of the stock market determines that the analysis and forecast of stock is an important applied field for the data mining technology. Many scholars and companies are engaged in the research and application of data mining in stock analysis and forecast.This paper starts from the financial statements analysis of listing company. By adopting data mining technology into real use and analyzing those financial statements of listing company, it predicts the stock price trend of this company. The main research is included as following:(1) Analyzing the superiority of kernel covering algorithm on the basis of comparing other algorithms. Then according to stock price forecast model, do correspondent experiments for the sake of reaching the purpose of forecasting earning margin per share of the listing company's stock in the incoming year. The result shows that even without many samples kernel covering algorithm has far more forecasting accuracy rate than the traditional one (such as BP algorithm) and a little bit higher than the forecast accurate rate of SVM (Support Vector Machine) , which proves the applicability of kernel covering algorithm in stock classified forecast. Moreover, it also provides a brand-new method and thought for stock forecasting.