Keywords :
|
Generalized feed forward, Multi Layer Preceptron, fuzzy neural network (Adaptive Neuro-fuzzy inference systems, forecasting, stock price.
|
Abstract :
|
Stock price forecasting is a prominent phenomenon for investors and other financial information users. Indeed, stock price prediction has been regarded as an interesting yet challenging process in the advanced world of business. Various economical or non economical factors have effected on stock market behavior, hence stock price forecasting is recognized as one of the most complicated subjects in business. In past, statistic-based methods were suggested to solve this problem. In recent decade, nonlinear based fuzzy time series models methods, artificial neural networks, fuzzy neural networks and combined prediction models have been recommended. There have been permanent debates about different methods of forecasting precision among those authors who selected artificial intelligence in forecasting. Therefore comparative analyses are crucial. This study by applying neural networks and minimizing stock price forecasting error designs and provides stock price forecasting model, in comparison to combined artificial neural networks technique. Results show that the combined neural network model forecasts more appropriately and is faster with higher estimation capability in stock price forecasting in relation to single neural networks.
|