Forecasting mutual fund net asset values using artificial neural networks

The purpose of this study is to forecast net asset values of Turkish mutual funds using Artificial Neural Networks (ANN) method. In order to forecast net asset values of 38 mutual funds (19 A type and 19 B type), 6 macro economic variables are used. These variables are Active Bond Interest Rate, USD/TL Exchange Rate, ISE National 100 Index, M2 Money Supply, Industrial Production Index and Wholesale Price Index.The forecasting period consists of monthly closing prices for these variables in the period of January 2001-December 2008. In this study, net asset values of mutual funds have been forecasted within the frame of both Artificial Neural Networks and regression model and forecasting performances of the results obtained from two methods have been compared.Analysis results reveal that ANN method is capable of forecasting net asset values of mutual funds at a very low error level. Furthermore, forecasting capability of ANN is compared with regression method and ANN method seems to outperform regression method.