Index Construction for Pattern Mining over Probabilistic Event Streams

2020 Fiscal Year Final Research Report

• Project/Area Number: 19K21530
• Project/Area Number (Other): 18H06461 (2018)
• Research Category: Grant-in-Aid for Research Activity Start-up
• Allocation Type: Multi-year Fund (2019); Single-year Grants (2018)
• Review Section: 1001:Information science, computer engineering, and related fields
• Research Institution: Nagoya University
• Principal Investigator: Sugiura Kento 名古屋大学, 情報学研究科, 特任助教 (10821663)
• Project Period (FY): 2018-08-24 – 2021-03-31
• Keywords: 確率的イベントストリーム / ストリーム処理 / 索引構造

Outline of Final Research Achievements

In this research, we aimed to develop an index structure that can assist analytical processing such as pattern mining for probabilistic event streams. We proposed a method for efficiently calculating the appropriate probability of occurrence of patterns described by regular expressions. We also surveyed the state-of-the-art lock-free indexes and re-implemented some of them, and clarified their performance characteristics and unsolved problems. These results provide a basis for developing an index structure for probabilistic event streams.

Free Research Field

データベース・データ工学

Academic Significance and Societal Importance of the Research Achievements

機械学習技術は近年大きな注目を集めた一方で，それによって得られる不確実なデータの処理方法は未だ発展途上である．本課題の目指すところは入力データの不確実性を考慮した最終的な分析結果の取得及びその不確実性の算出であり，不確実なデータからの妥当かつ実用的な結果の取得を補助するという意義がある．また，最新の索引構造の再現実装をとおして元論文では述べられていない知見も得ており，新たな未解決課題の提示を行ったという点でも意義がある．