Abstract:
Time changes of return, inefficiency studies performed and presence of effective factors on share return rate are caused development modern and intelligent methods in estimation and evaluation of share return in stock companies. Aim of this research is prediction of return using financial variables with artificial neural network approach. Therefore, the statistical population of this study includes 120 listed companies in Tehran stock securities during 2005 to 2017. Independent variables in this research are market variables (Earning quality, free cash flow) and dependent variable is share return. The obtained outputs from estimation of the artificial neural networks and results obtained from estimation, using of this method with evaluation scales concerning random amount and comparing it with adjusted R, we found that there is meaningful relation between the associated variables and return. However, such network has the least error than other networks.
Machine summary:
The obtained outputs from estimation of the artificial neural networks and results obtained from estimation, using of this method with evaluation scales con-cerning random amount and comparing it with adjusted R, we found that there is meaningful relation between the associated variables and return.
Ku-mar and Ravi [9], 128 articles studied in relation with prediction of bankruptcy of bank and compa-nies, and resulted that artificial neural network method was better than many methods combined sys-tems can have been better performance by combination advantages and differences of these methods.
/ Vol. 4, Issue 2, (2019) / Advances in mathematical finance and applications / [105] / Prediction the Return Fluctuations with Artificial Neural Networks’ Approach 3 Proposed Methodology Statistic society studied in this research consists of all listed companies in Tehran stock Exchange in 2007 to 2017.
Artificial neural networks are shown good performance as a modern method in model-ing and forecasting nonlinear and nonpermanent time series of process which there is no clear solution and relation to their exact identification and description.
For predicting return using research variables with feed forward neural network approach that is generally called multilayer perceptron networks (MLP) is used, for training the above neural network make use of error back propagation learning rule.
/ Vol. 4, Issue 2, (2019) / Advances in mathematical finance and applications / [109] / 5 Results and Estimates The methods of artificial neural networks with use of sensitivity analysis will determine the corre-lation of the input variables (independent) and output (dependent).