Mining association rules from consumer product safety cases based on text classification

Consumer product safety is closely related to people's health. The associations between impact factors of consumer product safety can be discovered by mining association rules from consumer product safety cases. Text classification based on improved term weighting method is proposed for the discretization of the unstructured text of consumer product safety cases. An improved candidate generation function based on Apriori algorithm is presented for generating all frequent itemsets. To verify the proposed methods, two experiments are conducted. The experimental results combined with the expert evaluation indicate that our methods can effectively mining association rules from consumer product safety cases. The presented solution for association rule mining has been successfully applied to the "Impact Factors of Consumer Product Quality & Safety System".