CompareD: a text mining solution for research & innovation proposals evaluation

In the context of the ISA² programme, the CompareD platform has been developed to support the evaluation of research proposals by public research funding agencies. 

With a user-friendly interface, the web-based platform allows evaluators of Research and Innovation (R&I) applications to compare incoming proposals with a corpus of grants and other relevant documents. Users can upload pieces of text or files in all the EU languages and the platform applies an automatic translation into English, which is the language of the corpus of reference data. The text similarity computation provides many options for users to fine-tune the matching of the documents, including a more detailed comparison for detection of plagiarism. The corpus of data in CompareD currently includes the database of R&I EU-funded projects (CORDIS) and other projects funded by European Agencies.

The platform also provides a full workflow of task assignment, deadlines, and audit activities between two types of user profiles: the reviewer and the manager. Both profiles can check the history of previous comparisons with associated results and statistics. An additional function also allows for the automatic extraction of metadata from PDF documents by using AI.

How could the CompareD platform be useful?

CompareD brings more evidence to funding decisions by evaluators, helps detect plagiarism and contributes to avoiding duplication of research, helps to increase the quality and originality of publicly funded research projects and facilitate the sharing of public funding research data in Europe. The platform can also assist the selection of panel of reviewers, since it allows the matching of proposals with Scopus data from Elsevier.

Click here for more information on CompareD. If you are a European public or private funding agency for R&I and interested in testing the platform you can get in touch at JRC-COMPARED-SUPPORT@ec.europa.eu.