數(shù)據(jù)挖掘_Wikilens Data Set(WikiLens數(shù)據(jù)集)

1、Wikilens Data Set(WikiLens數(shù)據(jù)集)數(shù)據(jù)摘要：WikiLens was a generalized collaborative recommender system that allowed its community to define item types (e.g. beer) and categories (e.g. microbrews, pale ales, stouts), and then rate and get recommendations for items.It was taken offline in 2009 due to lack of

2、system maintenance and support.This data set was extracted in February 2008.中文關(guān)鍵詞：WikiLens,推薦系統(tǒng),項目類型,類別,英文關(guān)鍵詞：WikiLens,recommender system,item types,categories,數(shù)據(jù)格式：TEXT數(shù)據(jù)用途：Information ProcessingClassification數(shù)據(jù)詳細(xì)介紹：Wikilens Data SetsWikiLens was a generalized collaborative recommender system that

3、allowed itscommunity to define item types (e.g. beer) and categories (e.g. microbrews, pale ales, stouts), and then rate and get recommendations for items.It was taken offline in 2009 due to lack of system maintenance and support.This directory contains a dump.txt.gz for this WikiLens instance.This

4、file is a gzip-ed output of a mysqldump command with the 'latin1'charset, after suitable erasing of private data.The intent is for this dump to have all data you could get byspidering the site.The easiest way to see the data is to install MySQL (see create a database, and loadthis dump file,

5、 e.g.zcat dump.txt.gz | mysql -uuser -ppassword -Dmy_databaseOtherwise, the text dump is human-readable, so it is possible to write tools to parse it. I wouldn't recommend it.The dump has the following tables:category - Map items to categorieschefmoz - Cache of Chefmoz import data for the Restau

6、rant category. This is EMPTY because otherwise it is very large.link - Cache of wiki page linkslogging - Log of actions taken on the wiki. This is EMPTY for privacy. member - EMPTY.nonempty - Cache of page ids of pages that have some content.page - Page data.page_urn - Mapping of pages to URNs for r

7、atings.pref - User preferences. This is EMPTY for privacy.rating - Ratings of URNs (often pages, mapped through page_urn). NOTE: rateepage is a URN id, not a page id.recent - Cache of page ids of pages recently changed.session - Cache of user sessions for the wiki. This is EMPTY for privacy. urn - U

8、RN (Universal Resource Identifier) ids.user - EMPTY.version - Page data for every version of a page.These tables are mostly the same as PhpWiki 1.3.9 (see). The new tables are category,page_urn, rating, and logging.* WARNING *The easiest mistake to make while looking at the data is to join the ratee

9、page field of the rating table and the id table of page.rateepage is a page id, right? NOT SO. The rating.rateepage field is actually the id of a URN, NOT a page. That field name has not been changed to something reflecting URN simply due to lack of time to do it correctly (including database migrat

10、ion upgrades).Look carefully at the example queries below to see how to use the various fields.* WARNING * Example queriesThe words "item" and "ratee" (the object of a rating action) are used synonymously below. Similarly for "user" and "rater".Here are some e

11、xample queries to get data:1. Select all ratings. Columns are- Ratee (item) page id- Ratee (item) page name (truncated to 25 characters)- Rater page id- Rater page name- URN id- Rating value- Rating timestampselect p.id, left(p.pagename, 25), r.raterpage,rp.pagename as rater, r.rateepage, r.ratingva

12、lue as rat, r.tstamp from page p, page rp, rating r, page_urn pu, urn uwhere pu.pagename = p.pagename and pu.urn = u.urn and r.raterpage = rp.idand r.rateepage = u.idorder by p.pagename2. Select all ratings of an item called "Book_Foo"select p.id, left(p.pagename, 25), r.raterpage,rp.pagen

13、ame as rater, r.rateepage, r.ratingvalue as rat from page p, page rp, rating r, page_urn pu, urn uwhere pu.pagename = p.pagename and pu.urn = u.urn and r.raterpage = rp.idand r.rateepage = u.id and p.pagename like 'Book_Foo' order by p.pagename3. Select all ratings of a user called "Use

14、r_Bar"select p.id, left(p.pagename, 25), r.raterpage,rp.pagename as rater, r.rateepage, r.ratingvalue as rat from page p, page rp, rating r, page_urn pu, urn uwhere pu.pagename = p.pagename and pu.urn = u.urn and r.raterpage = rp.idand r.rateepage = u.id and rp.pagename like 'User_Bar'

15、order by rp.pagename4. Select number of things in the "Book" category:select count(*) cnt from category c, page pwhere c.category = p.id and p.pagename = 'Book'5. The number of items in any categoryselect count(*) from category6. The number of usersselect count(*) cntfrom category

16、c, page pwhere c.category = p.id and p.pagename = 'User'7. The number of ratingsselect count(*) from rating8. The number of ratings per monthselect left(tstamp, 6) as yearmonth, count(*) from ratinggroup by yearmonthorder by yearmonth asc9. Pages per category for every categoryselect pagename, count(*) cntfrom category c, page pwhere c.category = p.idgroup by categoryorder by cnt desc10. Ratings per category for every categoryselect left(cp.pagename, 30) cat, cou