Presentation on theme: "IR CW2–09 Webpage summarization 01/13 Can webpage summarization improve the search engine user experience? By Barney Staddon."— Presentation transcript:
IR CW2–09 Webpage summarization 01/13 Can webpage summarization improve the search engine user experience? By Barney Staddon
IR CW2–09 Webpage summarization 02/13 The need for webpage summarization Types of summary The approaches The problems The solutions (Do they work?) Conclusions
IR CW2–09 Webpage summarization 03/13 The need for webpage summarization Vertical listing is used by most search engines If a summary is provided, shouldn’t it be useful? “Only 1 in 4 user-queries is initially successful ” Microsoft (2009) A good summary could avoid ‘dead’ visits Fewer ‘dead’ visits makes a better experience
IR CW2–09 Webpage summarization 04/13 Types of summary Extract or abstract? Query-relevant or generic?
IR CW2–09 Webpage summarization 05/13 The approaches Content-based summarization: Target webpage Summary
IR CW2–09 Webpage summarization 06/13 The approaches Context-based summarization: (source: Amitay & Paris, 2000)
IR CW2–09 Webpage summarization 07/13 Content problems Webpage text Is there a pre-authored summary available? What text is important and relevant? Are words, phrases or sentences extracted? Is it good quality? Summary
IR CW2–09 Webpage summarization 08/13 Context problems Content text Where does the context come from? Is there enough context text and is it relevant? Are words, phrases,or sentences extracted? Is it good quality? Summary
IR CW2–09 Webpage summarization 09/13 The solutions Open Directory Project HTML parser Term Frequency – Inverse Document Frequency Lexical chains (disambiguation) Sentence segmentation Contextual linking Contextual query data
IR CW2–09 Webpage summarization 10/13 Bing ‘hover’ (www.lsbu.ac.uk)
IR CW2–09 Webpage summarization 11/13 Bing ‘hover’ (www.football.co.uk)
IR CW2–09 Webpage summarization 12/13 Conclusions Content-based summarization - More likely to find good quality pre-authored summary - Random extracts can be more like a preview - More space is useful Context-based summarization - Only as good as the search engine it’s linked to - Requires greater processing power Can webpage summarization improve the search engine user experience? Yes! Preview/summary/excerpt/snippet – representative & viewable. Is cohesion necessary? A better way to present?
IR CW2–09 Webpage summarization 13/13 Can webpage summarization improve the search engine user experience? Questions References: Microsoft. (2009). Bing: New Features Relevant to Webmasters. [Online]. Available from: publishers-released.aspx [Accessed 03/12/09] publishers-released.aspx Amitay, E. and Paris, E. (2000). Automatically Summarising Web Sites - Is There A Way Around It? [Online]. Available from: amitay.pdf?key1=354816&key2= &coll=GUIDE&dl=GUIDE&CFID= &CFTOKEN= [Accessed 03/12/09]http://delivery.acm.org/ /360000/354816/p173- amitay.pdf?key1=354816&key2= &coll=GUIDE&dl=GUIDE&CFID= &CFTOKEN= [Accessed 03/12/09]