Sequential Index Structure for Content-Based Retrieval

Data mining applied to databases of data sequences generates a number of sequential patterns, which often require additional processing. The post-processing usually consists in searching the source databases for data sequences which contain a given sequential pattern or a part of it. This type of content-based querying is not well supported by RDBMSs, since the traditional optimization techniques are focused on exact-match querying. In this paper, we introduce a new bitmap-oriented index structure, which efficiently optimizes content-based queries on dense databases of data sequences. Our experiments show a significant improvement over traditional database accessing methods.