Predictionof Stock Price by Fuzzy Artificial Neural Network Approach Using Markovand ARIMA Model

Stock price prediction isone of the ways that can be used to utilize investment opportunities and better allocate resources. Firstly, by providing the necessary warnings, companies can be made alert to falling stock prices so that they can take appropriate measures with respect to these warnings, and secondly, investors and creditors can distinct favorable investment opportunities fromunfavorableonesand invest their resources in the proper cases.One of the most important ways to increase the accuracy of companies’ stock price prediction is to use new data mining techniques forprediction. Given this fact, there is growing interest in the theoretical development of free dynamic intelligent systems which are based on empirical data.The backup vector machine is one of these dynamic systems that transmits the knowledge or law hidden in the data to the algorithm and network structure by processing empirical data. Considering the above issues, the main issue of this research is to investigate and compare the predictive power of companies stock prices using Fuzzy Neural Network compared to ARMA and ARIMA models. In this regard, in this research, 770 years-companieswere chosen as the final sample. Of the 770 samples selected, 650 were reported as test data and 120 as test data.To ensure model validity, the sample was divided into a training sample and a test sample.

Keywords:

Stock Price , Fuzzy Artificial Neural Network , MarkovModel , ARIMA Model