Application of bitmap index to information retrieval

To solve this problem, we devised the HS-bitmap index, which is hierarchically comprised of compressed data of summary bits. A summary bit in an upper matrix is obtained by logical OR of the n bits in its corresponding lower matrix. Let n denote the summary unit. Summary units are constant and determined by an appropriate rule in the document-term matrix. The hierarchical summary bits provide bitmap indices with efficient bit operations, usage of space for large sparse bitmaps, and partial decompression. To evaluate the efficiency of the HS-bitmap index, we implemented it on top of PostgreSQL, compared the query performance of the HS-bitmap index with that of the inverted index, and found that the HS-bitmap index performs better especially when queries are long. We developed the HS-bitmap index for efficient information retrieval. The HS-bitmap index is a hierarchical document-term matrix: the original document-term matrix is called the leaf matrix and an upper matrix is the summary of...

• Type: Conference
• Year: 2008
• Where: WWW
• Authors: Kengo Fujioka, Yukio Uematsu, Makoto Onizuka
• Keywords: Hierarchical Document-term Matrix, HS-bitmap Index, Internet Technology, Lower Matrix, WWW 2008