Database compression with data mining methods

  • Authors:
  • Chien-Le Goh;Kazuki Aisaka;Masahiko Tsukamoto;Shojiro Nishio

  • Affiliations:
  • Osaka Univ., Osaka, Japan;Osaka Univ., Osaka, Japan;Osaka Univ., Osaka, Japan;Osaka Univ., Osaka, Japan

  • Venue:
  • Information organization and databases
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

Despite the drop in prices, storage cost is still a major cost factor in large scale database applications, such as data warehouses. Data compression is needed to reduce the cost. Many data compression techniques have been proposed and the issue of database compression has been discussed. Conventional data compression techniques require that compressed data be decompressed before read operations or write operations can be carried out. As a result, it is not practical to compress databases in active use using the conventional data compression techniques. In this chapter, we propose a database compression technique which needs only partial decompression for read operations and no decompression for write operations. It is suitable for databases in active use and can be used to compress data in relational databases. The proposed technique finds rules in a relational database using the Apriori Algorithm and store data using rules to achieve high compression ratios. The rules are in turn stored in a deductive database to enable easy data access.