Learnability of Interestingness with Semantic Similarity and Reasoning in the Knowledge Base of Decision Support Systems

Related Work

The author has developed question answering over knowledge-based (KBQA) as a tool to retrieve the answer from a large scale semi-structured knowledge base (Jin, Luo, Gao, Tang, & Yuan, 2019). The author enumerates the challenges locating the key entities and their relations from the knowledge base from the input question. The other problem is to concentrate on the search space of the final answer after linking the key entities to the underlying knowledge base. To rank the retrieved subgraphs after mapping the query graph onto the knowledge base also becomes a challenge. These challenges tend to evaluate the semantic similarity between the question and node edge subgraph containing the answer is a hurdle to overcome.

Jiang et al. (2019) explores semantic technologies to be applied to open dataset searches to improve the usability and efficiency of the search for the ever-increasing collection of open data.

The contributions of the approach are:

• •

  Applying the ontology-based semantic approach with hybrid indexing to open data domain to improve search quality

• •

  Automatic linking to save time and reduce the need for required domain expertise

• •

  Multi-language support, i.e., the search is independent of the natural language used in the dataset description.

Due to the lack of standard measure (Jiang et al., 2019) proposes two measures Concept Dataset Similarity Vector (CDSV) and Concept Concept Similarity Matrix (CCSM).