A data mining approach for efficient selection bitmap join index (Q601267)

Summary: The amount of information in a data warehouse tends to be extremely large and queries may involve several complex join and aggregates operations at the same time. To improve performance of these queries, database administrators often use indices. However, selection of an optimal set of indices is a very hard task because of the exponential number of attributes candidates that can be used in the selection process. To deal with this problem, we propose a data mining pruning approach based on maximal frequent itemsets representing candidate attributes for the index selection process. The main particularity of our pruning approach, compared to the existing ones, is that it uses other parameters than the frequency constraint, and respect monotony and anti-monotony properties. A greedy algorithm is proposed in order to select indices using a subset of attribute candidates. These indices minimise the query processing cost and satisfy the storage constraint. We validate our proposed algorithm using an experimental evaluation.