The Study Of Trading Behavior Of Investors Based On K-means Algorithm And Decision Tree Model

With the development of Internet technology,more and more information has been left by users on the Internet.It is of great significance for business decision-making and theoretical research by the application of data mining technology to find the inherent law between behavior and reveal the potential value information behind it.In the field of Internet finance,the trading behavior of investors also contains a lot of information.In order to exploring the characteristics of stock investors' transaction behavior,this paper studies the law of trading behavior and stock returns with the research object of stock investors' trading behavior.The main contents of this paper include five chapters.in the first chapter,we introduce the background of data mining,the process of data mining and the content of mining.Then it introduces the research of data mining in the field of finance and user behavior.Finally,on the basis of the research of various scholars,we put forward the research direction and significance of this paper,and briefly introduce the main contents of this paper.In the second chapter,we first collect data which is related to the investor from the related network,and we preprocess the data,define and quantify the investment trading behavior of users.We analyze the whole trading behavior on the profit and loss situation,the operating style,transaction number,frequency and the stock selecting.It is found that the overall transaction record is more profitable than the loss,the overall transaction is conservative,and the stock selection is not good.Most transactions tend to be short or medium-term investments and tend to operate at low frequencies by the analysis.In the third chapter,we take the medium and short term transaction records as the research object.Through the K-means algorithm,we divide them into four categories,and analyze the behavior characteristics of each category,with the income of each category.and we find that the medium and short term profit situation is greatly influenced by the transaction style in the whole transaction,and its operating frequency is less influential.In the fourth chapter,the investor is the research object to define and transform the index,and then the investor's index is discretized by K-means algorithm,then the decision tree model is constructed by ID3 algorithm.After the model training and pruning,the rules between the investor's behavior characteristics and the income are obtained,and the investor's stock selection ability is most important to the income of the investors.In addition,the timing ability,operation style and trading frequency also have some influence on the investor's income.On the basis of this rule,we put forward some investment suggestions for investors.In the fifth chapter,we summarize the content of the whole article,expound the shortcomings of this paper,and discusses the improvement of the next model.At the same time,the development of the research direction and the significance of the research are prospected.