Knowledge Inference Based Fuzzy Temporal Associative Classification And Its Applications On System Of Food Safety Management

Association rules are used to find the links between different products among transactions database, these rules reflect the buying behaviors of customers. In recent years, the association rules based on time constraint become a hot research. In other words, mining the transactions which have the time-sequence can get the temporal sequence patterns of user behaviors. In particular, a problem under the influence of ambiguity need to be solved in temporal association rules. In this paper, we research on a series problem of temporal association rules and fuzzy temporal association rules, including the definitions of support and confidence, temporal association rules algorithm, pruning methods of temporal association rules and inferences of temporal association rules.Firstly, we introduce the classic algorithm of Apriori and the association rules algorithm based on FP-tree. On basis of that, temporal logic and temporal database are introduced firstly. Besides that, we propose the temporal associative algorithm and temporal associative classification algorithm. In order to improve the efficiency of algorithm, we propose a new temporal associative classification algorithm based on GR-tree search mechanism and three new pruning methods. Through test, the efficiency of the improved temporal associative classification algorithm was greatly improved.On the basis of temporal association rules and temporal associative classification algorithm, we introduce the fuzzy logic and redefine the support and confidence of association rules so that propose the fuzzy temporal associative classification algorithm. In order to solve the low efficiency caused by scanning database repeatedly and adding the fuzzy logic, we propose a new idea which divides the temporal sequent database into some non-overlapping sub-database to search respectively. After comparing fuzzy temporal associative classification algorithm with temporal associative classification algorithm, efficiency and accuracy of the algorithm is greatly improved because relevant fuzzy temporal association rules are increased. Finally, we apply the fuzzy temporal associative classification algorithm in inference of food safety management and mining the fuzzy temporal association rules for the food safety officer assessment of food safety laws and regulations. On the basis of that, we get the fuzzy temporal association rules and eventually provide the basis of decision making for regulations of food safety management.