Mining the Future: Predicting Itemsets´ Support of Association Rules Mining

This paper proposes a novel research dimension in the field of data mining, which is mining the future data before its arrival, or in other words: predicting association rules ahead before the arrival of the data. To achieve that, we need only predict the itemsets´ support, upon which association rules could be easily produced. A time series analysis approach (MFTP) is proposed to perform itemsets´ support prediction task. The proposed technique outperforms other prediction techniques for short history. The conducted performance study showed good prediction accuracy and response time. Thus, we provide a new tool to provide more information in the decision support field