Gold Price Forecasting Based on RBF Neural Network and Hybrid Fuzzy Clustering Algorithm

This paper predicts good price based on RBF neural network employing hybrid fuzzy clustering algorithm. PCA technique has been used to integrate the 6 parameter dependent sub-variables of each TI (Technical Indicators, include MA, ROC, BIAS, D, K), which has been originated from the gold price before, and the results act as input. By employing a new hybrid fuzzy clustering algorithm, which is proposed by Antonios and George [10], K-Mean clustering algorithm and RBE algorithm, the predictions of price are yielded for each interval-n model. n refers to the number of predictions achieved by 1 operation. The most important conclusion indicates that the hybrid fuzzy clustering algorithm is superior to the general RBF central vector selecting algorithm mentioned above, in the aspects of MSE, P-Accuracy Rate and ROC.