ASSIST Project

The ASSIST project is the result of the community call stage of the ASSERT project which provided funding opportunities to investigate the benefits of text mining in two case studies within the social science disciplines. This includes a review of the requirements gathering stage in order to advise future projects in this area and the development of high profile exemplars demonstrating how text mining solutions can solve, in part at least, major challenges facing e-Researchers across all domains.

Two projects have been funded, a brief description of each is provided below. As the projects develop we will expand this site to include more information and provide links to the developing exemplars.

Duration: February 2008 - July 2009

Project Team

Principal Investigator: Sophia Ananiadou
Co-Investigators: John McNaught, NaCTeM and James Thomas, Evidence for Policy and Practice Information and Co-ordinating Centre and Peter Halfpenny, National Centre for e-Social Science
Project Team (NaCTeM): Davy Weissenbacher and Brian Rea

Case Study 1: UK Educational Evidence Portal

This project is with the Evidence for Policy and Practice Information and Co-ordinating Centre (specifically its user involvement team) to work with NaCTeM to develop and evaluate an innovative search engine — using text mining — for a portal of education evidence, relevant to education practitioners and policy-makers. This project will be a high profile exemplar of the utility of text mining in the social sciences, with application beyond the single case described here.

Case Study 2: Frame Analysis of Media

This project is with the ESRC National Centre for e-Social Science (NCeSS) in collaboration with the ESRC Centre for Research on Socio-Cultural Change (CRESC) to work with NaCTeM to develop and evaluate an innovative tool for analysing news texts to investigate how they are framed to shape the perceptions or opinions of the information's recipient. An outcome of the project will be an evaluation of the applicability of text mining tools, initially developed for quantitative data analysis, to improve qualitative analysis.

Overview of the ASSIST classification system

Queries are expanded via different source of information. The results of the research are clustered automatically and can be corrected by the user.

Progress of the project

A beta version of a demonstrator expanding on the ASSERT system has been adapted to our corpus and has been deployed. The system processes two different corpus. The first corpus is composed of 4900 textual documents extracted from the Lexis Nexis newspapers database with a query defined by our NCeSS partner. The second corpus has been provided by our EPPI partner. It is composed of 1300 'PDF', 'Microsoft Word' and 'XHTML' documents coming from selected academic education web sites. In collaboration with our partner, we currently in the process of a qualitative evaluation of this demonstrator. This evaluation would lead us to enrich its design and improve the access of the document content for the user of this specialized search engine.

Demonstrators (beta version)

This general presentation shows the overall architecture of the ASSIST search engine and details the functionalities with screen shoot illustrations (pdf).

This video clip demonstrates the main functionalities of the ASSIST search engine for the Frame Analysis of Media.
Demonstrator for Educational Evidence Portal

Dissemination

Presentations

The presentations on the ASSIST Project were given in the "Text Mining and the Social Sciences" Workshop in conjunction with the 4th International Conference on e-Social Science. [Slides here]
A comparison between the ASSIST system and the CAQDAS tools was presented during the panel on the MeCCSA Conference (2009) [Slides here]

Publications

Ananiadou, S., Thompson, P., Thomas, J., Mu., T. , Oliver, S., Rickinson, M., Sasaki, Y., Weissenbacher, D. and McNaught, J., Supporting the Education Evidence Portal via Text Mining. Philosophical Transcations of the Royal Society A 2010, 368(1925), 3829-3844. [Article]
Weissenbacher D., Rea B., Ananiadou S., Text Mining: beyond the CAQDAS tools? Paper presented at the panel on Innovations in Methods in Media and Communication Studies at the Media, Communication and Cultural Studies Association (MeCCSA) 2009, Bradford [Short paper]
Weissenbacher D., Rea B., Ananiadou S., Are the CAQDAS and the Text Mining Software Competitors? Fourth International Conference on Interdisciplinary Social Sciences 2009, Athens [Abstract]
Ananiadou S., Weissenbacher D., Rea B., Pieri E., Lin Y., Vis F., Procter R., Halfpenny P., Supporting Frame Analysis using Text Mining. 5th International Conference on e-Social Science 2009, Cologne [long paper]
Weissenbacher D., Pieri E., Ananiadou S., Rea B., Vis F., Lin Y., Procter R., Halfpenny P., ASSIST: un moteur de recherche spécialisé pour l’analyse des cadres d’expériences Conférence sur le Traitement Automatique des Langues Naturelles 2009, Senlis [Short paper]