Comparative Evaluation of XML Information Retrieval Systems: by Saadia Malik, Andrew Trotman, Mounia Lalmas, Norbert Fuhr

By Saadia Malik, Andrew Trotman, Mounia Lalmas, Norbert Fuhr (auth.), Norbert Fuhr, Mounia Lalmas, Andrew Trotman (eds.)

This ebook constitutes the completely refereed post-proceedings of the fifth foreign Workshop of the Initiative for the assessment of XML Retrieval, INEX 2006, held at Dagstuhl citadel, Germany, in December 2006.

The forty nine revised complete papers offered have been rigorously chosen for presentation on the workshop and went via a next around of cautious reviewing and revision. The papers are geared up in topical sections on technique, and seven extra tracks on ad-hoc, average language processing, heterogeneous assortment, multimedia, interactive, use case, in addition to record mining.

Show description

Read Online or Download Comparative Evaluation of XML Information Retrieval Systems: 5th International Workshop of the Initiative for the Evaluation of XML Retrieval, INEX 2006, Dagstuhl Castle, Germany, December 17-20, 2006, Revised and Selected Papers PDF

Best international conferences and symposiums books

Database Theory — ICDT’99: 7th International Conference Jerusalem, Israel, January 10–12, 1999 Proceedings

Databaseresearchisa? eldofcomputersciencewheretheorymeetsapplications. Many thoughts and strategies, that have been considered as problems with theoretical curiosity while before everything proposed, at the moment are incorporated in applied database structures and comparable items. Examples abound within the ? elds of database layout, question languages, question optimization, concurrency regulate, statistical databases, etc.

Interactive Distributed Multimedia Systems and Telecommunication Services: 5th International Workshop, IDMS'98 Oslo, Norway, September 8–11, 1998 Proceedings

This ebook constitutes the refereed court cases of the fifth overseas Workshop on Interactive allotted Multimedia platforms and Telecommunication providers, IDMS'98, held in Oslo, Norway, in September 1998. The 23 revised complete papers offered have been conscientiously chosen from a complete of sixty eight submissions.

Graph-Theoretic Concepts in Computer Science: 31st International Workshop, WG 2005, Metz, France, June 23-25, 2005, Revised Selected Papers

This publication constitutes the completely refereed post-proceedings of the thirty first overseas Workshop on Graph-Theoretic options in computing device technology, WG 2005, held in Metz, France in June 2005. The 38 revised complete papers provided including 2 invited papers have been conscientiously chosen from one hundred twenty five submissions.

Additional resources for Comparative Evaluation of XML Information Retrieval Systems: 5th International Workshop of the Initiative for the Evaluation of XML Retrieval, INEX 2006, Dagstuhl Castle, Germany, December 17-20, 2006, Revised and Selected Papers

Sample text

In Figure 3, this method selects the node: a. Full Recall-Base (Full RB): For comparison reasons, we also use the full recall-base where all relevant nodes, including overlapping nodes, appear in the ideal set. From the sample tree in Figure 3, this method selects all relevant nodes: a, b, c, d, e, f, g and h. Fig. 3. An example XML tree. Relevant nodes are indicated by filled circles. The specificity score is shown next to each relevant node. The Leaf and Article Nodes methods represent the two extremes of all possible selection mechanisms.

07) and the Article Node (all nodes here have depth of 1) methods. Comparing the nxCG measures at various cutoffs, we see that as the cutoff is increased, the results produced by the Standard, S. , Leaf Node and P. of Leaf Nodes methods tend to agree more with each other’s conclusions, and tend to disagree with the results of the Article Node method. The same can be said for the MAnxCG measures. Between the nxCG and MAnxCG measures, we can see that the latter is more resilient to changes in the ideal recall-base at lower cutoffs.

Exhaustivity value [7]) in INEX 2006; and (3) the probability that the user goes from one element in the list to a target (BEP). For the best in context retrieval task, this probability is defined as s(x, b), as defined in Equation 14, for any BEP b. e. we only know that a random user sees the BEP with probability s(x, b) if presented the element x in the list. With these settings, a ranking only made of BEPs will obtain a constant precision of 1 for all recall levels. The performance slowly decreases when returned elements are further away from the BEPs, and reach 0 when returned elements are not in relevant articles.

Download PDF sample

Rated 4.73 of 5 – based on 14 votes