Temperature prediction based on fuzzy clustering and fuzzy rules interpolation techniques
Yu‐Chuan Chang, Shyi‐Ming Chen · 2009
In this paper, we present a new method to deal with temperature prediction based on fuzzy clustering and fuzzy rules interpolation techniques. First, the proposed method constructs fuzzy rules from training samples based on the fuzzy C-Means clustering algorithm, where each fuzzy rule corresponds to a cluster and the linguistic terms appearing in the fuzzy rules are represented by triangular fuzzy sets. Then, it performs fuzzy inference based on the multiple fuzzy rules interpolation scheme, where it calculates the weight of each fuzzy rule with respect to the input observation based on the defuzzified values of triangular fuzzy sets. Finally, it uses the weight of each fuzzy rule to calculate the forecasted output. We also apply the proposed method to handle the temperature prediction problem. The experimental result shows that the proposed method gets higher average forecasting accuracy rates than Chen and Hwang's method.