FRIS

The project

The FRIS Enricher enriches the metadata of the FRIS Flanders Research Portal (researchportal.be/) by creating new keywords for each publication or project.

The software exploits various natural language processing techniques such as TextRank (an extractive, unsupervised text summarization), a modified version of TF-IDF (a numerical statistic that compares keyword candidates to keywords of other papers) and automated synonym extraction of manually created keywords.

Each individual technique assigns a score to its keyword candidates. The more techniques decide for one specific keyword, the higher the score gets. The highest scoring keywords of each publication get sent back to FRIS. Our service also applies language checking on the input data and finds a pdf link for publications (if it is freely available).

Some researchers forget to fill in keywords or fill in too few when they upload their research to the FRIS Portal which makes it difficult to find publications or projects with the search engine. This software was developed to enrich the metadata with more keywords in order to improve the searching experience on the FRIS Research Portal. By adding more specific keywords, it becomes easier to find a specific publication or project.