In this paper we will briefly describe the approaches taken by Berkeley for the main GeoCLEF 2007 tasks (Mono and Bilingual retrieval). This year we used only a single system in ...
This paper presents a quantitative performance analysis of two different approaches to the lemmatization of the Czech text data. The first one is based on manually prepared diction...
For this year's Image CLEF Photo Retrieval task, we have prepared 5 submission runs to help us assess the effectiveness of 1) image content-based retrieval, and 2) textbased ...
Abstract. Information Retrieval is a supporting technique which underpins a broad range of content-based applications including retrieval, filtering, summarisation, browsing, clas...
We propose a weakly-supervised approach for extracting class attributes from structured text available within Web documents. The overall precision of the extracted attributes is a...