Univrsity Of Tehran Press Financial Research Journal 1024-8153 19 2 2017 07 23 Investigating Performance of Bayesian and Levenberg-Marquardt Neural Network in Comparison Classical Models in Stock Price Forecasting Investigating Performance of Bayesian and Levenberg-Marquardt Neural Network in Comparison Classical Models in Stock Price Forecasting 299 318 64813 10.22059/jfr.2017.214203.1006264 FA Hossein Fakhari Associate Prof. of Accounting, University of Mazandaran, Mazandaran, Iran 0000-0003-4192-6582 Mohammad Valipour Khatir Assistant Prof. in Industrial Management, University of Mazandaran, Mazandaran, Iran Maedeh Mousavi MSc. in Accounting, University of Mazandaran, Mazandaran, Iran Journal Article 2016 08 20 Accurate forecasting of stock prices according to high volatility and inherent risk of stock market is a major concern of investors and financial analysts, hence applying novel approaches to predict the stock priceisan inevitable necessity. Accordingly, the purpose of this research is to compare the performance of forecasting models such as neural network with classical model and introducing appropriate model to forecast tomorrow stock price. The daily market prices data and financial indicator have been used as input variables for designing neural network model and daily closing price data set as input variable for designing ARIMA and also tomorrow's closing price is considered as output variable from 2011to2014. The results show that the Bayesian neural network represents less error sand higher Predictive power than the ARIMA model. The findings indicate the efficiency of Bayesian neural network incapture short-term investment opportunities and also can help investors to choose the appropriate portfolio and to obtain more returns. Accurate forecasting of stock prices according to high volatility and inherent risk of stock market is a major concern of investors and financial analysts, hence applying novel approaches to predict the stock priceisan inevitable necessity. Accordingly, the purpose of this research is to compare the performance of forecasting models such as neural network with classical model and introducing appropriate model to forecast tomorrow stock price. The daily market prices data and financial indicator have been used as input variables for designing neural network model and daily closing price data set as input variable for designing ARIMA and also tomorrow's closing price is considered as output variable from 2011to2014. The results show that the Bayesian neural network represents less error sand higher Predictive power than the ARIMA model. The findings indicate the efficiency of Bayesian neural network incapture short-term investment opportunities and also can help investors to choose the appropriate portfolio and to obtain more returns. https://jfr.ut.ac.ir/article_64813_6fcc1ede3600391ae0070c857adc17d5.pdf