MiRABIT: A new algorithm for mining association rules

Mining a sales transaction database is a very expensive task in terms of the processing time. Depending on the business branch of the company some characteristics of the database are different. The algorithms currently used for mining association rules optimize this task for databases with supermarket characteristics. The MiRABIT algorithm that is presented in this work optimizes the mining task for databases with clothing store characteristics. Comparative tests demonstrate the efficiency of the proposed algorithm in relation to the standard algorithm Apriori in a clothing store database.