Research on improving information quality on the supply side by "reverse" information recommendation

2022 Fiscal Year Final Research Report

• Project/Area Number: 19K12114
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 61030:Intelligent informatics-related
• Research Institution: University of Tsukuba
• Principal Investigator: Chen Hanxiong 筑波大学, システム情報系, 准教授 (60251047)
• Co-Investigator(Kenkyū-buntansha): 古瀬 一隆 白鴎大学, 経営学部, 教授 (10291288)
• Project Period (FY): 2019-04-01 – 2023-03-31
• Keywords: 情報推薦 / クエリ属性改良 / 逆ランク / 近傍検索 / 検索アルゴリズム

Outline of Final Research Achievements

Information recommendation provides users with efficient decision-making support. The research so far has been based on the premise that "users = consumers", and the main goal has been, for example, to recommend products and information to customers. This research regards information suppliers and producers as users, and based on the degree of matching with information users and consumers, provides information (products, services, information, etc.) that is immediately useful for quality improvement to suppliers. In this research, we developed various neighborhood/similarity search algorithms to approach this NP-hard problem. Furthermore, we devised a unique indexing method to solve the problem that spatial indexing is not effective for high-dimensional vectors. We also developed a prototype mobile application that simulates product improvement by recommending information to suppliers.

Free Research Field

情報学

Academic Significance and Societal Importance of the Research Achievements

商用データ、クラウド(crowd)コンピューティングで正当な評価データ生成を期待する一方、スパムデータの大量生成によって妨害も可能である。近年、SNSで”いいね”を買う政治家や自動生成アカウントで敵味方の発信にリスポンス・反応し、世論操作するような行為が多く見られた。「逆」情報推薦による供給側の情報品質向上を目的とする本研究で開発されたアルゴリズムはこのような場面に迅速に対応して発信情報を推薦するができ、本研究成果の波及効果の１つにあげたい。
また、迷惑メールデータ、ネットワーク侵入検知データを利用して脆弱要件を検出し、改良アドバイスの作成への応用が期待される。