Modeling virtual knowledge graphs using relevant news data by NLP methods for business analysis
Muhammad Arslan, Christophe Cruz · 2022
Companies aiming at online data sources for extracting business insights often experience the issue of managing different collections of datasets. It is because online business-related data is growing over time. This increasing data demands more storage. But saving the irrelevant data will cost more to companies. To overcome the problems of data integration and scalability, the idea of Virtual Knowledge Graphs (VKGs) is exploited. Moreover, text processing methods based on Natural Language Processing (NLP) are utilized for improving the identification of relevant business news data. The contribution of this work is twofold that is presented using a business case study. First, we have applied the concept of VKG to expand the company database. The expansion is required for adding more business-related data from the news articles to improve the process of identifying its relevance as per the end-user's interests. Second, we have proposed the text relevance resolution matrix and its working is shown through NLP techniques for identifying relevant news data. Populating the modeled VKG with relevant news data will save database space and prevent manual labor. Moreover, it will also help business analysts to understand business data temporally, spatially, and thematically.