ARTAR: Temporal association rule mining algorithm based on attribute reduction

Focusing on the defects existing in classical Association Rules Mining algorithms, we explore a new way of mining temporal association rules with the time-property in dataset. By combining with the Rough Set Theory and parallel computing technology, a temporal association rules mining algorithm based on attribute reduction - ARTAR is put forward to deal with high dimensional data. The Attributes Reduction of Rough Set Theory is utilized to reduce the number of data items in each transaction. In order to avoid frequent itemsets threshold setting problem, the Apriori integrating with Top-k theory is adopted. Meanwhile the Time-weighted factor is introduced as time constraints. The experiment analysises demonstrated that the proposed algorithm RTAR has higher efficiency and fits to temporal association rule mining in High-dimensional data.