Deepening BWT for massive data processing

2023 Fiscal Year Final Research Report

• Project/Area Number: 19K20213
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 60010:Theory of informatics-related
• Research Institution: Kyushu Institute of Technology
• Principal Investigator: I Tomohiro 九州工業大学, 大学院情報工学研究院, 准教授 (20773360)
• Project Period (FY): 2019-04-01 – 2024-03-31
• Keywords: 文字列処理 / BW変換 / 圧縮文字列処理

Outline of Final Research Achievements

The Burrows-Wheeler Transform (BWT) of a string is obtained by sorting each character in the string with its subsequent suffix, which has been used for data compression and compressed data processing. In this project we obtained the following results: (1) We simplified the index based on Run-length BWT (RLBWT) and improved its throughput for direct construction. (2) We proposed a practical algorithm for converting RLBWT to LZ77. (3) We proposed a BWT-based index for palindrome pattern matching. (4) We proposed an efficient algorithm to construct BWT-based indexes for parameterized pattern matching.

Free Research Field

文字列処理

Academic Significance and Societal Importance of the Research Achievements

データ処理において，データをどのように表現するかは処理の効率に大きく関わる最重要かつ根源的な問題である．圧縮のためのデータ変換手法として提案されたBurrows-Wheeler変換（BW変換）は，後の研究によりデータ処理において様々な利点を有していることが明らかになっている．本研究は，BW変換文字列を連長圧縮した領域で動作するアルゴリズムや一般化文字列照合におけるBW変換の応用技術の発展に寄与した．