Using snippets in text summarization: A comparative study and an application

Automatic text summarization consists of automatically creating a summary of one or more texts. As for Web pages, unfortunately classical techniques cannot be applied in presence of dynamic contents. In this paper, we propose the adoption of snippets -i.e., page excerpts provided together with user query results by search engines- as a text summarization technique. The study is conducted along two directions: comparing the proposed approach with a classical text summarization technique and (ii) assessing whether snippet summarization can be successfully applied to contextual advertising. On the one hand, comparative experiments show that the proposed approach has performances similar to those obtained by using the selected classical technique. On the other hand, the adoption of snippets as text summarization technique in contextual advertising show that the performances are quite satisfactory.