Fuzzy Inference Algorithm based on Quantitative Association Rules

In order to develop a data mining system to extract the fuzzy inference rules from the data, in this paper a fuzzy inference algorithm based on quantitative association rule (FI-QAR) is proposed. First, a discretization algorithm based on an improved clustering for each dimension data is adopted, and then the quantitative results are represented in the form of a Nominal variables matrix to compute the support and confidence level in the Apriori algorithm for quantitative association rules mining. On the basis of this, the quantitative association fuzzy rules are reconstructed by combing with TS fuzzy model to realize fuzzy inference, which can be applied to predict the output class and precise output. Experiment results demonstrated that the proposed algorithm is feasible and practical.