Home Aims & Objectives NaCTeM Services Text Mining Tools Text Mining Infrastructures Text Mining for Biodiversity Resources Terms & Conditions FAQ News & Events People Projects Current Projects AIRC EPHOR Mental Health Past Projects ADVISES Arabic WordNet ASSIST ASSERT AstraZeneca Project Automated screening for systematic reviews BBC Big Mechanism BOOTStrep Bott and Co. CheTA Clinical Trials COPIOUS DECA eScholar EMPATHY Europe PMC FixRep FLaReNet Graphene HSE Lloyds Infectious Diseases INTUTE ISHER KISTI Pathway META-NET Mining for Public Health Mining the History of Medicine MMPathIC NCS TOX ONDEX OpenMinTeD OSSMETER Pacific Life Re PathText/
Refine SLiM Thalia Turing Project Publications Community External Collaboration Vacancies Teaching & Tuorials Feedback How to find us Site Map Search

eScholar project

Overview

Manchester eScholar is a search facility at the University of Manchester that gives researchers access to scholarly work produced by individuals associated with the university. NaCTeM, in collaboration with the University of Manchester Library (UML), will enrich the current faceted search capabilities of eScholar by customising, adapting and combining existing text mining tools and algorithms to foster the discovery of interdisciplinary links.

The advancement of new interdisciplinary research is reliant on identifying potential synergies between the work of different groups within the university. Often, researchers in different schools or departments may not be aware of potential overlaps between their respective research and therefore collaborative opportunities may be missed. One way of discovering potential research links is to examine the papers and articles produced by different groups in order to identify possible commonalities in previously reported research. This project aims to automate the process of discovering such links through the application of advanced text mining techniques.

Techniques

The project will involve:

• keyword extraction, similar to TerMine, to discover important terminology related to a query;
• named entity recognition to foster semantic search, e.g., looking for “chemical:lead” should return a narrower, more focused set of results than the more general and ambiguous query “lead”;
• topic clustering to discover diverse groups of semantically related documents across various domains and dimensions/facets.

The proposed text mining techniques have been previously proven in other services such as Europe PubMed Central, in which NaCTeM introduced semantic querying, and search systems developed by NaCTeM, including the clinical trials faceted search engine and a news-based search engine developed in the context of the ISHER project.

Project Team

Principal Investigator: Prof. Sophia Ananiadou