Grey relational grades and neural networks : empirical evidence on vice funds

Abstract

This research examines time-series predictability of Vice Funds Indices through the Grey Relational Analysis (GRA), and also applies three types of Artificial Neural Networks (ANN) model, namely, Backpropagation Perception Network (BPN), Recurrent Neural Network (RNN), and Radial Basis Function Neural Network (RBFNN) to capture nonlinear tendencies of Vice Funds indices. The study finds that among the three ANN models, BPN has the best predicting power. When the data is separated into 10%, 33% and 50% testing data sets to test the proficiency of the available forecasting information in the timeseries of the predictors, the predictive power of the BPN model again dominated the findings 60% of the
time. Traders, investors and fund manager can rely on BPN predicting power with large or even small data set. Nevertheless, the result also suggests the predicting power of both RNN and RBFNN model with smaller data sets. Overall, it is suggested that traders and fund managers have stronger chance of achieving more accurate forecasting using the BPN model in Vice Funds indices. Findings of this research have policy implications in the creation of forecasting and investing strategies by examining models that minimize errors in predicting Vice Funds indices.