Abstract

The current research paper aims at constructing a model based on financial ratios for predicting the companies’ current and future return. To investigate the financial ratios’ ability in elaborating the current stock return and prediction of the future stock return, the decision tree was applied. Based thereon, a collection of logical constraints in the format of an algorithm featuring a tree structure is used for the prediction and recognition of the outcomes of an event. So, models were formulated by four decision tree algorithms, including CHAID, ECHAID, QUEST, and CRT, based on 70% of the study data and the results obtained from an examination thereof were compared in 30% of the data residues in terms of such scales as accuracy, sensitivity, and specificity. The data from 317 companies accepted to the securities exchange market for the time span from 2002 to 2014 were used in the present study. The obtained results indicated that CRT and ECHAID algorithms deliver the best performance in elaborating the current return and the algorithm CHAID offers the best performance in predicting the future stock return. Also, models demonstrated greater power in elaborating the current stock return than the future stock return. Because the models were found unreliable in their abilities in both of the states, the assumption suggesting the existence of an elaborative relationship between the proposed financial ratios and the current and future stock return variations was rejected