Earnings per Share forecasting models using MLP and RBF Neural Networks on the listed firms in Tehran Stock Exchange (TSE)

Document Type : Original Article

Authors

10.22034/iaar.2014.104401

Abstract

Earnings per share (EPS) are one of the important financial ratios that is considered by manager, investors and financial analysts. It is usually used in investment decisions, profitability evaluation, profit risk, and stock price estimation. Therefore, EPS forecasting is an important and an attractive task for manager and investors. This research proposes a model for forecasting earnings per share using a multi-layered perceptron(MLP) neural network and radial basic functions(RBF) neural network and compares estimation accuracy of 2 models using performance criteria.For this purpose, we used 630 listed firms in Tehran Stock Exchange (TSE) in the period of 2003-2009. The results show that the MLP model is significantly more accurate than the RBF model.

Keywords

Accounting and Auditing Research
Volume 6, Issue 21
April 2014
Pages 100-115

Files

Share

How to cite

Statistics