عنوان مقاله [English]
نویسندگان [English]چکیده [English]
Prediction in financial affairs especially in securities is highly important.
Investors make wide evaluations while investing on stocks. One of the main factors considered by investors during their investments is to earn returns. In such circumstances, a suitable prediction model for stock returns will cause the allocation of optimal resources and efficiency in capital market which are important issues individually and nationally. Recent article addresses the way of predicting stock returns in Tehran Stock Exchange by using Arbitrage multiple regression model and artificial neural networks. The variables of the research includes 971 samples of four daily macro-economic variables namely TSE Dividend and Price Index (TEDPIX), gold prices, currency exchange rate (Rial/$) and the amount of transactions between Iranian calendar years 1381 and 1385 (2002-2006).
To process Arbitrage pricing model, multi-factor regression and to process artificial neural networks (ANNs), Perceptron architecture model with two hidden layers and back-propagation algorithm with sigmoid conversion functions are applied.
To assess the performance of both models, mean absolute deviation (MAD), mean square error (MSE), mean absolute percentage error (MAPE) and root mean square error (RMSE) are utilized.
The findings show the success of both models in predicting cash return index and Tehran Stock Exchange prices as well as the superiority of artificial neural network over Arbitrage multiple model.