Constructing Corpus for Query-Oriented XML Text Summarization

XML Retrieval is becoming the focus study of the field of Information Retrieval and Database. Summarization of the results which come from the XML search engines will alleviate the read burden of user´s. However, as the basis of this study, the construction of the query-oriented XML text summarization corpus has not yet received enough attention. In this paper, we introduce our works on constructing this kind of corpus, including the selection of topics and XML elements/documents, construction process and the feature of the constructed corpus. Up to now, the corpus has 25 English query topics, including 422 elements for summarization, and 32 Chinese topics which including 402 elements. For each topic, 4 pieces of extracted summaries and 4 pieces of generated summaries are made manually by 4 experts.