Predicting Stock Market Dividends Using Ranking and Data Mining Techniques

Abstract

Data mining is one of the growing sciences and is very suitable for the analysis of databases. Data mining is used in many sciences, such as business intelligence, shopping cart analysis, and medicine. The main data mining algorithms are 4 categories, that 2 of their main categories are attribute ranking and classification algorithms. So far, no research has been presented regarding the use of ranking in classification algorithms. In this research, we present a method for predicting dividend market price using ranking and data mining techniques. First, a new method for classifying data has been presented, then ranking algorithms has been used as the input of this method. Afterwards, by implementing the above approach on 10 databases and obtaining the accuracy of each model according to the model inputs (attribute ranking algorithm) and placing the accuracy obtained in each stage in the data envelopment analysis model, we have evaluated and ranked the attributes. We have used stock market data to make this research applicable and based on the approach, we have predicted the ratio of change in companies' dividends in 2015 according to the companies' data. The results indicated high accuracy of the proposed approach and its high speed.