CPS 49S Google: The Computer Science Within and its Impact on SocietyDiscussion FormatNoteIntroductionIntroduction (contd.)Discussion PointsSlide 7NextProximityAnchor textPageRankAssigned ReadingsCPS 49SGoogle: The Computer Science Within and its Impact on Society Shivnath BabuSpring 2007Discussion Format•Talk for 10-15 minutes–Give an overview•Give an outline the discussion points that you have come up with•Need a scribe–Volunteer?Note•Make it a habit to check the course web page daily for:–Updated notes (presentation, discussion report, and scribe notes)–Current and future schedule –AnnouncementsIntroduction•http://www.google.com/corporate/tech.html•Let us look at some numbers–From the paper–From searchenginewatch.comIntroduction (contd.)•Terms–HTML (look at the HTML for the class web page), Hypertext, link/hyperlink, inlink, outlink, anchor text, link graph –Search engine, meta search engine–Information retrieval, crawl, index •Terms that we will discuss later–PageRank, proximity, barrel, …Discussion Points•Motivation for Google–Human-maintained lists–Keyword matching only–Advertising --- conflict of interestDiscussion Points•Design Goal #1: High-quality search results–Hypertext–Proximity–PageRank•Design Goal #2: Good performance•Design Goal #3: Support for research activitiesNext•Problem: User types in a keyword-based search query. We have to (i) find result pages to answer this query, and (ii) rank these result pages–Proximity of terms–Anchor text–PageRankProximity•Of terms on a web page•E.g., phrases•E.g., “anatomy”, “search”, “anatomy search”•E.g., “google freshman seminar duke”•Other examples?Anchor text•Text around the link•Often accurate and concise description of page•May have terms that the page does not contain–“search engine”–Other examples?•Can return pages that have not been crawledPageRank•First cut: count inlinks•Basic idea --- “recursive” counting•Interpretation based on probability•DemoAssigned Readings•For Tue (1/23)–Continuation of the anatomy paper–Paper on “Taxonomy of Web
View Full Document
Unlocking...