3293

2892

كارشناسي ارشد

برنامه ريزي سيستمهاي اقتصادي

اصفهان: دانشگاه صنعتي اصفهان، دانشكده صنايع و سيستم ها

1385

دوازده، 146، [II] ص.: مصور، جدول، نمودار

رضا حجازي

اكبر توكلي

محمد رضا زماني

1396/09/05

كتابنامه

صنايع و سيستم ها

مهندسي صنايع و سيستم ها

ID2892

به فارسي انگليسي : قابل رويت در نسخه ديجيتالي

Abstract The artificial neural networks due to the ability of pattern recognition in widely area nowadays theyare one of the most widely used tools in forecasting and simulation scope Then they can be used inforecasting and stock market recognition where the stock price index is a representative total trend of itsmovement The forecasting of stock price index can authorize one of the important parameter of thedecision making of administration and private planners to advance economic purposes The stock priceindex is as a thermometer of national economy and a symbol of economic situation Regarding to signalsdivulge from itself economic planners financial managers and businesses react to it The stock exchangeindex is a tool for the evaluation and comparison performance of economic managers and administratorswhich it represents the sensitively of national economy and capital market to these decisions In this thesis Tehran Stock Exchange Index is predicted using a method of forecasting time series theimplementation of neural networks tools Regarding to the principles of artificial neural networks designaccomplishing numerous experiment on different models that they are developed based on alert each designparameters obtain final model N3 5 1 for time series forecasting with a sigmoid function as an activationfunction for both hidden and output layer and N5 4 1 for casual forecasting with a sigmoid and a linearfunction as a transfer function for hidden and output layers respectively Our empirical results demonstratethe superiority of ARIMA model to the artificial neural networks

رضا حجازي

اكبر توكلي

محمد رضا زماني