Using Semantic Relations between Keywords to Categorize Articles from Scientific Literature

The amount of digital data is growing exponentially, and it is time consuming for researchers and readers to locate relevant information. Hence, being up-to-date in a specific research field (or topic) is a tedious and complex task. Our final goal is to create an intelligent scientific search engine by taking semantic relations into account. Our approach described in this paper is the starting point of such a smart system. Semantic relations between keywords are extracted from scientific articles in order to later help in the process of browsing and searching for content in a meaningful scientific way. By computing the most correlated categories and domains inherited from the keywords, we are able to extract the correct meaning of these keywords in relation to the article's concept. Our approach achieves a precision of 0.92 for both categories and domains extraction and a recall of 0.89 and 0.96, respectively.