DAMS: Dynamic Association for View Materialization Based on Rule Mining Scheme

In data warehousing, view selection (VS) is an important aspect. Optimal VS needs to be materialized in order to minimize the overall data retrieval time. To support the same, performance metrics like memory constraints to save materialized views, query execution time, and query workloads needs to be addressed to reduce the overall retrieval time. As far as static view materialization (VM) is concerned, pre-computing strategies are required to execute the query workload prior to VM, but the approach is not scalable for small disk sizes. In the current era, the memory requirement is humongous to store pre-computed views in the materialized query table (MQT) that adds an overhead to view maintenance cost and disk sizes. To address the aforementioned issues, the authors propose a novel VM scheme DAMS. DAMS operates in three phases. In the first phase, the scheme chooses a materialized view in a dynamic and on-demand basis to reduce the query processing time. Then, in the second phase, a novel attribute selection algorithm is proposed based on association rule mining (ARM) technique in VS to address historical queries. It selects a candidate view from a pool of such views. As the number of queries is large, the proposed algorithm reduces the computational latency in fetching the view result. Finally, selected views are prioritized by grouping items as clusters set based on support and confidence metrics to speed up VM operations.