Poster: At the Virginia Tech Digital Libraries Research Laboratory, we are advancing research on several fronts to make ETDs more accessible and usable for students and researchers, both in English and across languages. As part of ongoing work to make ETDs available across languages, we are automatically producing concept maps from ETDs written in English, and then automatically translating them into Spanish using phrase lists automatically mined from ETD collections from Mexican universities. We are extending this work from English - Spanish to English - Chinese by mining translations of Chinese words and phrases in the computing and telecommunication domains. We are using text mining techniques to help identify more relevant ETDs for users from large ETD collections, thus resulting in increased precision and recall. We are also developing software to harvest ETDs from NDLTD, to extract specific information from ETDs (e.g. 'Future Work' , 'Problem Statement' etc.) and to recommend papers similar to users' interests.

Date

2008