Application Of Decision Tree And Neural Network In Stock Classification Prediction

With the development of economy, the stock market has gone deep into China’smarket economy in various fields. The stock market is seen as a " barometer " ofmarket economy, reflects a country’s overall economic situation. Since the stockmarket play an important role in economic activity, more and more people choose toinvest in stocks. However, stock market is a complex system, it is difficult to makeaccurate predictions with traditional methods. In order to manage the stock marketbetter, We urgently need to find a way to dig out useful information from massivedata, to help us find the relationships they exist from a lot of stock market data. Withthe development of data mining technology, people finally find such a tool-datamining technology, it can handle the stock market’s huge, cumbersome anddisorganized data. In this paper, using C5.0decision tree, BP neural network andRBF neural network algorithm to establish three categories classification model.In the establishment of classification model, choosing200of Chinese A-sharemarket listed company as samples, selecting14representative financial indicators ofseven categories, using SPSS Clementine modeling tools, with the method of decisiontree classification and neural network classification method to establish classificationmodel. In modeling, select80%of the sample as the training samples for establishingmodel; select20%of the remaining sample as the testing sample for testing theaccurate of the model. After modeling, compared various models, then choose the bestprediction method. With the eventual model, analyze company’s finacial condition toidentify common financial feature of good companies, provided assistances forinvestors in stock investment decisions.