Unit Trust Forecasting using Adaptive Neural Fuzzy Inference System: A Performance Comparison

AbstractUnit trust market is equally important as stock market as both are contributed significantly to nation's economic performance. Success in investing unit trust may also promises attractive benefits for investors. However, tasks to ensure successful prediction are highly complicated as many uncertainty and unpredictable factors involved. In this paper, the forecast ability of Net Asset Value (NAV) of three unit trust funds with Adaptive Neural Fuzzy Inference System (ANFIS) is examined. The objective of this study is to forecast NAV of three unit trust funds using ANFIS. Three unit trust funds were selected to model and forecast the NAV. One by four of input structure for each unit trust was defined prior to determining fuzzy rules in the fuzzy forecast. The experimental results indicate that the model successfully forecasts the NAV of the unit trust funds. The forecasting errors for the three funds were in the ranges of [-0.2461, 0.1], [-0.1384,0.08], and [-0.025,0.015]. The Pru Bond Fund recorded the least errors among the three funds. ANFIS offers a promising tool for economists and market players in dealing with forecasting NAV of unit trusts