Creation of Materialized Views Based on Neural Network Algorithm

Materialized view (MV) is one of many technical optimizations like index and fragmentation. This technique ensures a similar answer for a workload. In other word, it is used to prohibit the repetition of the same computes and reduce the execution time. In this paper, we will present an algorithm based on neural networks and business intelligent to avoid repeating the creation of MV process. This algorithm has two phases: The first phase is the learning of the logic between the arrived queries and the materialized views as the final solution. However, the second phase allows the prediction of a new solution associated to the new workload. In order to calculate the similarity between the vectors (query-attributes), we use the Jaccard Index. Finally, an experimental work is done to validate our approach.