An intelligent approach for mining knowledge graphs of online news

Automatic extraction of background knowledge of news articles, published over the web is a complex task and less explored by knowledge-mining researchers. Suppose there is no source of background or related information available on editorial, in this case, the reader needs to search it, which is a time-consuming process of manually going over possibly an in-definite number of knowledge sources to check its authenticity. This paper aims to resolve this issue of fetching associated information regarding any text by identifying the entities in text and fetching the related entities. This work has proposed a model that can automate this process by generating a semantic network to build a Knowledge Graph on any chunk of textual content. It identifies the entities and their corresponding relations and further expands them to extract related entities from non-trivial knowledge sources. Extracted entities and relations are further converted into a node(s) and edge(s) in an environment where knowledge is represented in a graph format to generate a Knowledge Graph. The mining of such a Knowledge Graph is a complex task by acquiring and integrating different information from heterogeneous sources. The resulting Knowledge Graph would be significant to be used as a knowledge base for future news. © 2021 Informa UK Limited, trading as Taylor & Francis Group.