In contrast to the closed LBD setting addressed in Chapters 5 and 6, open LBD adds another dimension to the discovery process by allowing that only domain 𝐶 is predetermined, thus facilitating the search for multiple potential candidates for the domain 𝐴 via intermediate domains 𝐵. To set the ground, the distinction between closed and open LBD is explained in Section 7.1. The first approach combining open and closed discovery is a concept-based approach presented in Section 7.2. The approach is illustrated by replicating Raynaud’s syndrome–Fish oil discovery through concept mapping, followed by the presentation of a semantic filtering approach to open discovery. Another approach, combining the text- and outlier-based open LBD, is the RaJoLink approach, presented in Section 7.3, which is illustrated in a concrete application in the autismcalcineurin domain pair. The chapter concludes with Section 7.4, which includes Python tutorials that allow experiment replicability and code reuse.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Semantic and Outlier-based Open Discovery

  • Nada Lavrač,
  • Bojan Cestnik,
  • Andrej Kastrin

摘要

In contrast to the closed LBD setting addressed in Chapters 5 and 6, open LBD adds another dimension to the discovery process by allowing that only domain 𝐶 is predetermined, thus facilitating the search for multiple potential candidates for the domain 𝐴 via intermediate domains 𝐵. To set the ground, the distinction between closed and open LBD is explained in Section 7.1. The first approach combining open and closed discovery is a concept-based approach presented in Section 7.2. The approach is illustrated by replicating Raynaud’s syndrome–Fish oil discovery through concept mapping, followed by the presentation of a semantic filtering approach to open discovery. Another approach, combining the text- and outlier-based open LBD, is the RaJoLink approach, presented in Section 7.3, which is illustrated in a concrete application in the autismcalcineurin domain pair. The chapter concludes with Section 7.4, which includes Python tutorials that allow experiment replicability and code reuse.