FIM-Frequent Itemset Mining was introduced in 1993 when Agrawal analysised data model supermarkets [1] and as a base to expand into the other problems in the area of data mining.In the study of the market, FIM in transaction database is to find the itemset often appear in the transaction. The FIM algorithm usually applied downward closure property [2] to increase the collective candidate pruning. Specifically, if there is an uncommon practice, the algorithm does not consider X files containing the candidate X, i.e. with a dataset containing n item and X contains k elements, the algorithm will not consider 2 (n-k )-2 sets containing X. However, FIM is only interested to buy or not buy items that are not interested in profit for each item. Therefore, the problem of high extraction useful collection is in place. We consider the example in Table 1 on sales data [3] to better understand the problem of exploitation of common practice and HUI mining problem.
đang được dịch, vui lòng đợi..
