Research In Stock Recommendation Based On Rough Set

Because of the rapid development of China’s economy, people’s living standard is raising, with growing capital. As one of the most important components in the financial market, stocks are becoming more and more active with people’s gaining awareness of financial investments. Stocks is one of the most important components of the financial market. How to make full use of the historical data in the stock information database to explore the orderliness of stock price has become extremely important. In recent years, amateurs use fundamental analysis, technical analysis and evolution analysis to predict the stock market, these methods are suitable for mature stock market. However, the history of Chinese stock market is quite short, and there are many retail private investors with little knowledge in finance, therefore, people need a more effective way to choose right stocks.Rough set theory is a mathematical tool to analyze and handle all kinds of imprecise and incomplete information, and it has become an important component in data mining research. In the premise of maintaining the same classification ability, knowledge acquisition of rough set theory conducts decision-making rules of the problem through attribute reduction. Due to the related data of profitable stocks cannot be comprehensive, rough set theory is more applicable in stock data mining.This paper focuses on attribute reduction based on the algorithm of attribute importance in rough set theory. First, we define the fluctuation of stock price as rate of change with data selected by relative performance of the requirements for data preprocessing and set the threshold value to meet the threshold of left stock data. Then we apply the incremental trend for discretization and turn stock time series into information system. With one trading day divided into two trading units, which are morning and afternoon, we look over the overall price change in both morning and afternoon. According to stock price and other attributes of the stock in both morning and afternoon, such as whether the stock price today becomes higher or lower and complex right price, get the rhythm of stock’s rise or fall. This solves the problem in inaccuracy of original stock attribute reduction by using high frequency trading with the probability of winning the way get the number of rules number further decreasing.On the basis of theoretical research, the simulation experiment is carried out by using MATLAB. We complete the implementation of the stock recommendation based on rough set with four steps, which are continuous attributes discretization, attribute reduction, rule extraction and rule interpretation. This journal applies the accurate historical stock transaction data to test the system of rough set with stocks, and achieved satisfying results. This paper does not only provide a prediction rule for the stock price but also help investors to choose the suitable stock for investment proposal.