Knowledge Sharing and Yahoo Answers: Everyone Knows Something
Lada A. Adamic and Jun Zhang and Eytan Bakshy and Mark S. Ackeerman
Knowledge Sharing and Yahoo Answers: Everyone Knows Something
International World Wide Web conference 2008
pp.665-674
PDFのある場所へのリンク
æ¦è¦
Yahoo Answerï¼æµ·å¤çYahoo!ç¥æµè¢ï¼ã®è§£æã
ã¨ã«ãããããã解æãã¦ããããçµå±ä½ãè¨ãããã®ãããããããªãã
ABSTRACT
Yahoo Answersï¼YAï¼ã¯åºãå¤æ§ãªquestion-answer forumã§ãtechnical knowledgeãå
±æããããã®åªä½ã¨ãã¦ã ãã§ãªããadviceãseekããããopinionsãgatherãããããããããªå¥½å¥å¿ãæºããããã®å ´æã¨ãã¦ãæ©è½ãã¦ããã
ãã®è«æã§ã¯YAã®ç¥èã®å
±æã®æ´»åï¼knowledge sharing activityï¼ã«ã¤ãã¦ã®ç解ãæ±ããã
forumã®ã«ãã´ãªã¼ã解æãããã®ä¸èº«ã®ç¹å¾´ã¨ã¦ã¼ã¶ã¼å士ã®interactionã®ãã¿ã¼ã³ã«ãã£ã¦ã¯ã©ã¹ã¿ãªã³ã°ããã
ããã¤ãã®ã«ãã´ãªã¼ã®interactionã¯expertise sharing forumã«ä¼¼ã¦ããããä»ã¯discussionãeveryday adviceãsupportãincorporateãã¦ããã
ã«ãã´ãªã¼ã®å¤æ§æ§ã«ããããããã幾人ãã®ã¦ã¼ã¶ã¼ã¯ç¹å®ã®topicã«ããããã¦focusãããã¨ãããã£ããä»ã¯ã«ãã´ãªã¼ãã¾ããã§åå ãã¦ããã
ãã®ãã¨ã«ããrelatedã«ãã´ãªã¼ãmapã§ããã ãã§ãªããã¦ã¼ã¶ã¼ã®interestã®entropyãcharacterizeã§ããã
factual expertiseãä¸å¿ã®ã«ãã´ãªã¼ã§ã®ã¿lower entropyã¯higher answer ratingã«correlateãããã¨ãããã£ãã
ä¸ããããã«ãã´ãªã¼ã®ä¸ã§ãç¹å®ã®answerãbest answerã«é¸ã°ãããã©ãããäºæ¸¬ããããã«ã¦ã¼ã¶ã¼attributeã¨answer characteristicãcombineããã
1. INTRODUCTION
knowledge exchange communityã§æ大ã®ãã®ã®ä¸ã¤ãYAã
ç¾å¨23Mã®è³ªåã解決ããã¦ããã
ãã誰ãä½ãç¥ã£ã¦ããã°ãYAã«ã¯ç¢ºå®ã«ãããshareããample opportunityãããã
knowledge sharingã¯ä¼çµ±çã«é£ããã£ãããYAã¯ç¤¾ä¼å
¨è¬ã®bootstrap knowledgeã®ã¡ã«ããºã ã¨æããcollective intelligenceãæä¾ãããã¨ã«ããéæããããã«æãããã
YAã§å
±æãããç¥èã¯ã¨ã¦ãå¹
åºãããä¸è¬çã«ãã¾ãdeepã§ã¯ãªãã
質åã¨åçã®å¤æ§æ§ãåçã®å¹
ãåçã®qualityãexamineãããããã«å¾ã£ã¦network and non-network analyzeãç¨ãã¦ã«ãã´ãªã¼ãåæãããã¨ã«ãããããã¤ãã¯technical expertise sharing forumã«ä¼¼ã¦ãã¦ãä»ã¯éã£ãdynamicsãæã£ã¦ãããã¨ãããã£ãã
ã¦ã¼ã¶ã¼ã®ã«ãã´ãªã¼ãã¾ããã åçãã¿ã¼ã³ã«ããç¥èã®åºãããè¨ãããã«entropyã®æ¦å¿µãå©ç¨ããã
ä½ãentropyãæã¤ãã¤ã¾ãé«ãfocusãæã¤ãã¨ã¯ç¹å®ã®ã«ãã´ãªã¼ã§ã®ãã¹ãåçã®å²åã«é¢ä¿ãããã¨ãããã£ããããã¯è³ªåãäºå®ãæ±ããã«ãã´ãªã¼ã§ã ãã
æå¾ã«åçã®qualityãexamineãã¦replierã¨answer attributeãã©ã®åçããã¹ãã«ãªãããããäºæ¸¬ããã®ã«ä½¿ãããã¨ãããã£ãã
2. PRIOR WORK
sharing knowledgeã¯å°ãªãã¨ã15年以ä¸ã¯research topicã§ãæè¿ã§ã¯ã¤ã³ã¿ã¼ãããã¹ã±ã¼ã«ã§ã®ãããé¢ç½ãã
Wikipediaãªã©ãå«ã¾ããã*1
ãã®åéã«ã¯ï¼ã¤ã®perspectiveãããã
ä¸ã¤ç®ã¯forumã®éããç¥ããã¨ã
Whittakerãã¯è¨å¤§ãªUsenet newsgroupã®ãã¼ã¿ã§general demographic patternãæããã«ãã解æãè¡ã£ã¦ããã*2
ããã«ã¯social network analysisãå©ç¨ããã¦ããã
ãã¨ãã°Kouã¨Zhangã¯asking-replying networkãæ²ç¤ºæ¿ã®ã·ã¹ãã ããæ§æãã¦äººã
ã®ãªã³ã©ã¤ã³ã§ã®interactionãpersonal interest spaceã«å¼·ãå½±é¿ãã¦ãããã¨ãç 究ããã*3
Fischerãã¨Turnerãã¯online interactionã®visualizationã®ç 究ãããã*4 *5
äºã¤ç®ã¯ã¦ã¼ã¶ã¼ã¬ãã«ã«focusããç 究ã
Wengerã¯éãå½¹å²ã®éè¦æ§ã¨ããããã³ãã¥ããã£ã®formationã¨continuationã«ã©ãå½±é¿ãã¦ããããè°è«ããã*6
Nonneckeã¨Preeceã¯lurkerã®æ¯ãèãã«ã¤ãã¦ç 究ããã*7
Donathã¯online forumã§ã¦ã¼ã¶ã¼ã®virtual identityãmineãã¦deceptionãdetectããæè¡ã調æ»ããã*8
æè¿ã§ã¯Welserããonline forumã§ã¦ã¼ã¶ã¼ã®ego-networkã"structural signature"ã¨ãã¦å©ç¨ã"discussion person"ã¨"answer person"ãidentifyã§ãããã¨ã«é¢ãã¦è°è«ããã*9
ä¸ã¤ç®ã¯threadã¨messageã¬ãã«ã«focusããç 究ã
Sackã¯visualizationã使ã£ã¦discussion threadã«ãããconversationã®ãã¿ã¼ã³ãæ§ã
ã§ãããã¨ã示ããã*10
Joyceã¨Krautã¯messeageã¬ãã«ã§å
容ã®åæãããnewcomerã®postã¨related responcesãå½¼ããåå ãç¶ãããå¦ãã«å½±é¿ãä¸ãã¦ãããã©ãããç 究ããã*11 *12
åã¤ç®ã¯ãªã人ã
ãonline communityã«åå ãï¼è²¢ç®ããã®ãã®ç 究ã
ããã¯ããã¦ãå°ããªdeta collectionã¨surveyã«ããè¡ããããä¾ãã°Lakhaniã¨von Hippel*13ãButlerã*14ã
our own workã§ããã種é¡ã®online forumã®ç 究ããã*15 *16ããããã®ç 究ã®goalã¯sharing knowledge and expertiseãinternet ageã§ãµãã¼ãããããã®ããè¯ãã·ã¹ãã ã¨online spaceããã¶ã¤ã³ãããã¨ã
ç 究ãéãã¦ãonline communityã«ããã大è¦æ¨¡ã®knowledge sharing and expertiseã®distributionã«é¢ãã¦æ¯è¼çå°ãããç¥ããã¦ããªããã¨ãç¥ã£ãã
YAã¯ãã®ç 究ã«ã¡ããã©è¯ããç¥ãéãã§ã¯äºã¤ã®ç 究ããè¡ããã¦ããªãã
Suãã¯YAã®answer ratingã使ã£ã¦ã¤ã³ã¿ã¼ãããã§ã®human reviewed dataã®qualityã«ã¤ãã¦ãã¹ãããã*17
Kimãã¯human codingã¨content analysisã使ã£ã¦ãã¹ãåçãé¸ã¶åºæºãç 究ããã*18
3. YAHOO ANSWERS AND DATA SET
YAã§ã®interactionã¯å
¨ã¦Q&Aã
質åã¯top-levelã®25ã®ã«ãã´ãªã¼ã¨ãlowerl levelã®1002ã®ã«ãã´ãªã¼ã«å±ããã
質åã«ã¯ãããããªç¨®é¡ãããã
"fact"ãæ±ãããã®ã
å©ããæ¯æ´ãæ±ãããã®ã
ç´ç²ã«discussionãè¡ãããå ´åãããã
ãã éè¦ãªã®ã¯ãããã®threadãYAã®ã«ã¼ã«ã«æ¯é
ããã¦ããã¨ãããã¨ã§ãåçã¯ï¼å以ä¸ã§ããªãããèªåèªèº«ã«ã¯åçã§ããªãããã®ç¹ã¯æããã«ä»ã®online systemã®thread interactionã¨ã¯éãã
YA activityãï¼ã¶æåéããã
8,452,337ã®åçã¨1,178,983ã®è³ªåã433,402ã®uniqueãªåçè
ã¨495,414ã®uniqueãªè³ªåè
ã質åã»åçãããããã¦ããã®ã¯211,372ã®ã¦ã¼ã¶ã¼ã ã£ãã
ãã®æ°åããã§ã«ã¦ã¼ã¶ã¼ã®å¤æ§æ§ã®ãã³ãã§ãããããããã®ã¦ã¼ã¶ã¼ã«ããå°ãªãpostæ°ã
4. CHARACTERIZING YA CATEGORIES
4.1 Basic characterics
ããããã®ã«ãã´ãªã¼ã¯factual informationãadvice seekingãsocial converasationã¾ãã¯discussionã®requestã®ããã¯ã¹ã ã¨æãã
ãããä¸èº«ãèªã¾ãã«æ£ç¢ºã«æ±ºããã®ã¯é£ããã®ã§ãavarage thread lengthï¼ï¼ãã¹ããããã®replyæ°ï¼ãaverage post lengthï¼åçãã©ãã ãé·ããï¼ã¨ãã£ãç¹å¾´ã観å¯ãããã¨ã§éæ¥çã«æ¨å®ããã
ããã°ã©ãã³ã°ãç§å¦ãç©çã¨ãã£ãtechnical subjectsã¯replyãå°ãªããããããã®replyã¯æ¯è¼çé·ããã¨ã観å¯ãããã
è¶
常ç¾è±¡ãªã©æ®éã®ç§å¦ããã¯é£ã³åºãã¦ãããã®ãæ±ã£ãsience subcategoryã¯å¤ãã®replyããã£ãã
ããä¸ã¤ã®æ¥µç«¯ãªcategoryã¯Jokes and Riddles categoryã§replyã¯çããæ°ã¯å¤ãã£ãã
discussion categoryã§ã¯æ®éã®é·ãã®replyãå¤ããã£ããWrestlingãªã©ã®ã¹ãã¼ãã®categoryããPhilosophyãReligionãPoliticsã¨ãã£ãcategoryãããã ã£ãã
ã¾ãMarriage & Divorce Marriageãããã¤ãã®è²å
ã«é¢ããcategoryãªã©ã®åã
ã®ä½é¨ãã¢ããã¤ã¹ãæ¢ããããªcategoryãããã ã£ãã
The Cats and Dogs categoryãããã ã£ãã
ä»ã®categoryéã§ã®ç¹å¾´ã®éãã¯asker/replier overlapã
ã¦ã¼ã¶ã¼ãæè¡çãªå°éç¥èãæ±ãããããªforumã§ã¯å¤§å¤æ°ã¯noviceã§ãããaskerã¨replierãããªãåºå¥ããããã¨ãäºæ³ãããã*19
ã¢ããã¤ã¹ãsupportãä¸å¿ã®forumã§ã¯ã¦ã¼ã¶ã¼ã¯æ±ããä¸ããããaskerããreplierã«ããªãããã ããã
discussion forumã§ã¯è³ªåã¨åçã¯ããããconversationãç¶ããããã®æ段ã§ããã
ããããã«technical categoryã®overlapãä½ããdiscussion forumãæãé«ãã£ãã®ã¯é©ãã¹ããã¨ã§ã¯ãªãã
ãã®ãã¨ã«é¢ãã¦ã¯Section 5.1ã§åã³æ±ãã
4.2 Cluster analysis of categories
ããããã®ã«ãã´ãªã¼ã«å¯¾ãã¦ããã¤ãã®aggregate measurementãããã
ããããã®ã«ãã´ãªã¼ã§ã®activityã¯Singles & Datingã§216,061ã®è³ªåãReligion and Spiritualityã§129,013ã®è³ªåãMathematicsã§48,624ã®è³ªåãDining Out in Switzerlandã§ã¯ãã£ãã®5ã®è³ªåã¨å¹
ããã£ãã
the most active categoriesï¼è³ªåã1000以ä¸ã®ã«ãã´ãªã¼ï¼ã®éåããthread lengthãcontent lengthãasker/replier overlapã®3ã¤ãããªããã¯ãã«ã«ããk-meansæ³ã§ã¯ã©ã¹ã¿ãªã³ã°ããã
thread lengthï¼ï¼è³ªåãããã®å¹³ååçæ°
content lengthï¼ï¼è³ªåãããã®å¹³ååçè
æ°
asker/replier overlapï¼ããããã®ã¦ã¼ã¶ã¼ã®è³ªåãåçã®é »åº¦ãè¦ç´ ã¨ããï¼ã¤ã®ãã¯ãã«ã®cosine similarity
解æããã®ã¯189ã®ã«ãã´ãªã¼ã§å
¨ä½ã®è³ªåã®91%ãå ããã
ã¯ã©ã¹ã¿ã¼ãï¼ã¤ã«ããã¨ãã«ç´æçã«ä¸çªæå³ãããããã ã£ãã
ä¸ã¤ç®ã®ã¯ã©ã¹ã¿ã¼ã¯discussion forumãããªã£ã¦ãã¦ã質åãåçãããã¦ã¼ã¶ã¼ãå¤ãããããããªã¹ãã¼ãã®ã«ãã´ãªã¼ã§ã¯åè
ã«ã¤ãã¦discussããããPoliticsã§ã¯squabble over partisanãè¡ããã¦ããããReligion & Spiritualityã§ã¯godã®æ¬è³ªã«ã¤ãã¦debateãããããã¦ãã¦ããã®ãããªåºæ¿çãªè³ªåã¯é·ãthread lengthã®åå ã«ãªã£ã¦ããã
äºã¤ç®ã®ã¯ã©ã¹ã¿ã¼ã¯ã人ã
ããããã¤ãæ£è§£ããã£ããå¯ä¸ã®facutual answerãåå¨ããªããããªè³ªåã§adviceyacommon-sense expertiseãæ¢ããæ±ãããããªã«ãã´ãªã¼ãããã§ã¯æ±ºå®çãªåçã¯ã»ã¨ãã©ãªããåæã«å¤ããã¢ããã¤ã¹ãããã®ã«é©åã«æããããã®ã§ãthreadsã¯é·ããªãå¾åãããã®ã§ããããFashionãBaby NamesãFast FoodãCatsãDogsãªã©ããµãã¾ããã
ä¸ã¤ç®ã®ã¯ã©ã¹ã¿ã¼ã¯å¤ãã®è³ªåãfactual answerãå«ãã人ã
ã¯è³ªåãåçãããthread lengthã¯çããªãå¾åããããBiologyãRepairsãProgrammingãªã©ãå«ã¾ããã
次ã®ã»ã¯ã·ã§ã³ã§è³ªåã¨åçã®dynamicsãããããã®ã¯ã©ã¹ã¿ã¼ã代表ããã«ãã´ãªã¼ã®ãããã¯ã¼ã¯æ§é ã®éãã解æãããã¨ã«ããããã«èª¿ã¹ãã
4.3 Network structure analysis
質åããã¦ã¼ã¶ã¼ãããã«åçããã¦ã¼ã¶ã¼ã«ã¤ãªããã¨ã«ããasker-replierã°ã©ããæ§æã§ãããããQAãããã¯ã¼ã¯ã¨å¼ã¶ãQAãããã¯ã¼ã¯ã®è§£æã¯non-network measureã§ã¯å®¹æã«captureã§ããªãéè¦ãªinterctionã®aspectã解æããããã®ã»ã¯ã·ã§ã³ã§ã¯ä¸ã¤ã®ã¯ã©ã¹ã¿ã¼ããå ¸åçãª3ã¤ã®ã«ãã´ãªã¼ãWrestlingãMarriage & Divorceï¼Marriage)ãProgramming & Designï¼Programmingï¼ã«ã¤ãã¦èª¿ã¹ãã
4.3.1 Detree distributions
ããããã®ã«ãã´ãªã¼ã§åºæ¬¡æ°ã¨å
¥æ¬¡æ°ã®cumulative distributionã観å¯ããã
ãã¹ã¦ã®ã«ãã´ãªã¼ã§ã¦ã¼ã¶ã¼ã®activity levelã®éãã観å¯ãããã
幾人ãã¯ããããåçããä»ã¯ï¼åãï¼åã§è³ªåãåçãããã¦ãã¾ããä¸æ¹æ¥µç«¯ãªä¾ã§ã¯ããããã®è³ªåã«åçããã¦ã¼ã¶ã¼ãããã
次ã«ããããã®ã«ãã´ãªã¼ã®éããè¦ããã
ï¼ã¤ã¨ãheavy tailed distributionã§ããããå
¥æ¬¡æ°ã®åå¸ã«ããã¦Marrigeã¨Wrestlingã¯ããbroaderã§ãããå°æ°ã®äººã
ã¯æ°åã®responseãï¼ã¶æéã«è²°ã£ã¦ããã
対ç
§çã«Programmingã§ã¯æã質åããã¦ããã¦ã¼ã¶ã¼ã§ãæ°åã®replyããè²°ã£ã¦ããªãã£ãã
ä¸è¬çã«ãYahoo answersã®forumã§ã¯åºæ¬¡æ°ã®åå¸ã¯broadã«ãªãå¾åãããã
Programingã§ã¯ããã¯consistentlyã«ä»ã®äººãå©ããèªåã¯å©ããå¿
è¦ã¨ããªããããªã¦ã¼ã¶ã¼ãåæ ãã¦ããã
Marriageã§ã¯ãããã¯regularlyã«adviceãæä¾ãããããªã¦ã¼ã¶ã¼ãããã¯Wrestlingã¨åãããã«discussionã好ããªã¦ã¼ã¶ã¼ã
ãã®roleã®separationã¯ãã¨ãã¦ã¼ã¶ã¼ãä¸ã¤ã®è³ªåãåçããã¦ããã¨è¦ãªãã¦ãæãããä¾ãã°Programmingã§ã¯è³ªåãããã¦ã¼ã¶ã¼ã®ç´57%ã¯æéä¸ä¸åº¦ãåçãã¦ããããåæ§ã«åçãããã¦ã¼ã¶ã¼ã®51%ã¯è³ªåãã¦ããªãã
4.3.2 Analysis of ego networks
Welserããonline forumã§ã®"answer person"ã¨"discussion person"ãego networkãã¿ããã¨ã«ããè¦åããããã¨ææ¡ãã¦ãã*20ã
ego networkã¯ã¦ã¼ã¶ã¼ã¨ç´æ¥é¢ä¿ã®ãã人ã¨ãã®äººãã¡ã®éã®ã¨ãã¸ã§ãããã
ï¼ã¤ã®ã«ãã´ãªã¼ããããããã©ã³ãã ã«100ã®ego networkãæ½åºãã¦æ¯ã¹ã¦ããã
Wrestlingã§ã¯highly activeã¦ã¼ã¶ã¼ã®é£äººã¯å½¼ãèªèº«highly connectedã§ããã¯å½¼ãã"discussion persons"ã§ãããã¨ã示ãã¦ããã
å対ã«Programmingã§ã¯æãactiveãªã¦ã¼ã¶ã¼ã¯ãhelpãã¦ããã¦ã¼ã¶ã¼ã¯ç¹ãã£ã¦ãããã"answer people"ã§ããã
4.3.3 Strongly connected components
ããããã®ã«ãã´ãªã¼ã®ãã¼ãæ°ãã¨ãã¸æ°ãå¹³å次æ°ãmutual edgeãæ大強é£çµæåã調ã¹æ¯ã¹ã¦ããã
Wrestlingã¯å¼·é£çµæåãæã¡ãæ¯è¼çå¤ãã®mutual edgeãæã£ã¦ãããcore social groupãã«ãã´ãªã¼å
ã«å½¢æããã¦ãããã¨ã示ãã¦ããã
Programmingã«ã¯ã»ã¨ãã©å¼·é£çµæåã¯ãªããreciprocal edgeã¯å®å
¨ã«ãªããããã¯"helpers"ã¨"askers"ã«å½¹å²ãåããã¦ããäºã«ããã¨ä¿¡ãã¦ããã
Marriageã¯ä¸éãmutual edgeã¯å°ãªããã¼ãã§ã¯ãªããæ大強é£çµæåã¯å°ãããåå¨ããã
次ã®sectionã§ãã詳ãã調ã¹ãã
4.3.4 Motif analysis
motif analysisã«ããç¹æã®social dynamicsã示ãinteractionã®small localãã¿ã¼ã³ãè¦ã¤ããããã
ããããã®ã«ãã´ãªã¼ã®å
¨ã¦ã®ãã©ã¤ã¢ãã«æ³¨ç®ãã¦å²åãæ°ããã©ã³ãã ãªãããã¯ã¼ã¯*21 *22 *23ã¨æ¯ã¹ãã
ã©ã³ãã ãããã¯ã¼ã¯ã«æ¯ã¹ã©ã®ã«ãã´ãªã¼ã§ãfeed forward loopãå¤ãã£ããããã¯ä¸äººãï¼äººãå©ãããã®ãã¡ã®ä¸äººãããä¸äººãå©ããã¨ããmotifãProgrammingã§å¤ãããã¯ãhigh levelã®expertiseããã¹ã¦ã®levelã®äººã
ãå©ãlower expertiseãããlowerãªäººãå©ãããã¨ããhelp-seeking online communityã§ããè¦ããã*24ç¹å¾´ã示ãã¦ããã
Wrestlingã¨Marriageã§ã¯fully reciprocal triadãå¤ãè¦ãããsymmetricãªé¢ä¿ãããã£ãããã®äºã¤ã®ã«ãã´ãªã¼ã®ããä¸ã¤ã®éè¦ãªtriadã¯ï¼äººã®mutual edgeãæã£ãã¦ã¼ã¶ã¼ï¼ãã¶ãforumã®regularï¼ã¨ãã®ã©ã¡ãã«ãåçãã¦ããã¦ã¼ã¶ã¼ï¼ãã¶ããã åã«è³ªåã«çããããã«åå ããã ãï¼ãããªãã
Programmingã§ã¯ããã¯å°ãªããããã¯ãäºãã«è³ªåã«çãããã£ã³ã¹ãæã£ã¦ããregularã¯ããactiveã§ãªãã¦ã¼ã¶ã¼ããåçãå¾ã¦ãããã¨ãimplyãã¦ããã
4.4 Expertise depth
質åã®depthã決ããããã«Programmingããã©ã³ãã ã«100ã®è³ªåãæ½åºãã5ã¤ã®levelã«rateããã
level3ã¯æ°å¹´ããã°ã©ã ã«ã¤ãã¦å¦ãã å¦çãããã®expertiseã
level4ã¯ããã®ããã°ã©ãã¼ãããã®expertiseã
Programmingã§ã¯level3ãè¶
ããexpertiseãå¿
è¦ã¨ãã質åã¯ï¼ï¼
ãããªãã£ãã
æçã«è¨ãã°ã質åã¯ã¨ã¦ãshallowã
5. EXPERTISE AND KNOWLEDGE ACROSS CATEGORIES
ãã®sectionã§ã¯ï¼ã¤ã®è¦ç¹ããYAã®åºããã«ã¤ãã¦è¨è¿°ããã
æåã¯ããã«ãã´ãªã¼ã§æ´»çºã«åçãã¦ãããªã¦ã¼ã¶ã¼ãä»ã®ã«ãã´ãªã¼ã§ãåæ§ã«ç·ã§ãããããªç¯å²ã«ã¤ãã¦èããã
次ã¯ã¦ã¼ã¶ã¼ã®entropyãããªãã¡å½¼ãã®åçãè½ã¡ãtopicã®å¹
ãè¨ãã
5.1 Relationships between categories
åçã®ãã¿ã¼ã³ã追跡ãããã¨ã«ããé¢ä¿ã®ããã«ãã´ãªã¼ãè¦ã¤ãåºããã¨ã¯ç°¡åã
ã«ãã´ãªã¼éã®è·é¢ããã«ãã´ãªã¼å
ã§åçãã¦ããã¦ã¼ã¶ã¼ã®éåã®cosine similarityãç¨ãã¦è¨ããé層ã¯ã©ã¹ã¿ãªã³ã°ãã¦ããã
ã³ã³ãã¥ã¼ã¿ã¼ä¸å¿ã®ã«ãã´ãªã¼ãComputer & InternetãConsumer ElectronicsãYahoo! ProductsãGames & Recreationãªã©ã¯åãã¯ã©ã¹ã¿ã¼ã«å«ã¾ããã
åãããã«Politics & Govermnmentã¯News & Eventsã¯ç¹ãã£ã¦ããã
Home & Gardenã¯Food & Drinkã¨ç¹ããFood & Drinkã¯Dining Outã¨ç¹ãããDining Outã¯Local Businessesã¨ç¹ãã£ã¦ããã
ãããã®ã«ãã´ãªã¼ãã¾ããã é¢ä¿ã¯ã¦ã¼ã¶ã¼å´ããã®èå³ã®focusã示åãã¦ããã
ããã«ãã´ãªã¼ã§åçããã¦ã¼ã¶ã¼ãåãã«ãã´ãªã¼ãããã¯ä»ã®ã©ã®ã«ãã´ãªã¼ã§åçããå¾åãããã®ãã質åã¨åçã®ãã¿ã¼ã³ãã調ã¹ãã
SportsãPoliticsãSociety & Cultureï¼Religionãå«ãï¼ã®ãããªdiscussionããã¡ãªtopicãæ±ãforumã§ã¯ã¦ã¼ã¶ã¼ã¯è³ªåã¨åçãåãforumã§è¡ããã¨ãå¤ããã¨ãobserveããã
Education & ReferenceãSience & Mathã®ãµãã«ãã´ãªã¼ã«è¦ããããããªäºå®ã«dominateãããtopicã§ã¯è³ªåãåçãããã¦ã¼ã¶ã¼ã¯å°ãªãã
è»ã¨è¼¸éã«é¢ãã¦åçãã¦ãã人ããä»ã®ã«ãã´ãªã¼ã§åçãã¦ãã¦ãè»ã«ã¤ãã¦è³ªåãã¦ãã人ã«ã»ã©ã«ã¯ä»ã®ã«ãã´ãªã¼ã§è³ªåããªãå¾åããããã¨ãããã£ãã
ã¹ãã¼ãã¨æ¿æ²»ã§ã®åçè
ã¯ç¾å®¹ã¨ã¹ã¿ã¤ã«ã«é¢ãã¦ã»ã¨ãã©è³ªåããªãã
åçããã«ãã´ãªã¼ã«é¢ä¿ãªããã¦ã¼ã¶ã¼ã¯ä¸æ§ã«Yahoo productsã«é¢ãã¦è³ªåãã¦ããã
åçããã«ãã´ãªã¼ã«é¢ãããHealthã§è³ªåããã¦ã¼ã¶ã¼ã¯å¤ãã£ãããããåçã®å¤ãã«ãã´ãªã¼ã§ããã£ãã
Faimly & relationshipsãHealthã¨åãããã§ãã£ãããrelationshipã«é¢ãã質åã¯æ¦ãã¦ä»ã®ã«ãã´ãªã¼ã§ã®åçã¨ã¯é¢ä¿ãªãã£ãã
technicalã«ãã´ãªã¼ã¨supportã«ãã´ãªã¼ã§ãé対称ãªé¢ä¿ããã£ããRelationshipsãHealthãParentingã§åçãã¦ããã¦ã¼ã¶ã¼ã¯Computers & Internetã§ã質åããããéã¯å°ãªãã
å°ãªãã¨ãæ°åã®ã¦ã¼ã¶ã¼ã¯å
¨ã¦ã®ã«ãã´ãªã¼ã§ã©ã³ãã ã«åçãã¦ãã訳ã§ã¯ãªãã®ã§åè¿°ã®ã«ãã´ãªã¼éã®é¢ä¿ã¯apparentã
ãªã®ã§ç¡æ°ã®topicã§knowedgeãå
±æããæ©ä¼ãããããã¦ã¼ã¶ã¼ã¯ç¯å²ãéå®ããå¾åãããã
5.2 User entropy
ç¹å®ã®topicã«ã©ãã ãã¦ã¼ã¶ã¼ãéä¸ãã¦ãããã®ææ¨ã¨ãã¦entropyã使ããéä¸ãã¦ããã°ä½ããªãã
ã«ãã´ãªã¼ã®é層ãåæ ããããããªentropyã®å®ç¾©ã«ããã
åçã®é »åº¦ã«ã¯ããªãå·®ããããtopicã«focusãã¦ããã®ããã ã»ã¨ãã©åçãã¦ããªãã®ããèæ
®ããªããã°ãªããªãã®ã§ã40以ä¸ã®åçããã41,266ã®ã¦ã¼ã¶ã¼ã«ãã¼ã£ã¦åæããã
ãããã®ã¦ã¼ã¶ã®ä¸ã§ãentropyã«ã¯ããªãå¹
ããã£ããããã¦ã¼ã¶ã¼ã¯èªããç¬ã®ãã¬ã¼ãã¼ã¨ç§°ããå
¨ã¦ã®åçã¯Dogãµãã«ãã´ãªã¼ã§ãããã«entropyã¯0ã ã£ãã
ä¸æ¹ã§40ã®è³ªåã25ããtop-levelã«ãã´ãªã¼ã®ãã¡17ã®ã«ãã´ãªã¼ã26ã®ãµãã«ãã´ãªã¼ã«æ£ãã°ã£ã¦ãããå½¼ã¯ã©ã®ã«ãã´ãªã¼ã§ãå
ã4ã¤ã®åçãããã¦ããããentropyã¯5.75ã ã£ãã
entropyã®åå¸ã¯é©ãããã¨ã«flatã ã£ããä½äººãã¯ã¨ã¦ãä½ãentropyã ããé«ãentropyãYAã®é層ã«ããéçã¾ã§æ¯è¼çæ®éã ã£ãã
ã¦ã¼ã¶ã¼ã®best answerã®ç¢ºçã調ã¹ãã¨ããããã®åå¸ã¯é対称ã§best answerç6-8%ã®ã¦ã¼ã¶ã¼ãä¸çªå¤ãã£ãã次ã®sectionã§ã¦ã¼ã¶ã¼ã®focusã¨best answerã«é¸ã°ãããã¨ã¨ã®é¢ä¿ãdetermineããããã«äºã¤ã®metricãcorrelateããã
5.3 Correlating focus to best answers
ç´æçã«åçãéãããç¯å²ã®ã«ãã´ãªã¼ã«focusãã¦ããå ´åãbest answerã«é¸ã°ããé »åº¦ã¯å¤§ããæ°ããããé¢ç½ããã¨ã«entropyã¨best answerã«é¸ã°ãã確çã«ã¯ç¸é¢ãç¡ãã£ãã
ããã¤ãã®discussion forumã§ã®åçã§ã¯ããã ããããå ´åã«ãã£ã¦ã¯ç¸é¢ãããã¨æå¾
ããã
ã«ãã´ãªã¼ã«ãéãããããfactual informationãæ±ãã«ãã´ãªã¼ã¯ä¸é¨ã§ãããã¨ã¯æ¢ã«å¦ãã ã
support forumã§ã¯best answerã¯æãempathyãããã¯caringãªadviceã§ãããã
discussion forumã§ã¯best answerã¯æã質åè
ã®æè¦ã«è³æãã¦ããåçã§ãããã
entertainment categoryã§ã¯æãwitãªåçãåã¤ã ããã
å
è¡ç 究ã§æ£ç¢ºãã詳細ããªã©ã®content valueã¯best answerã決ããè¦ç´ ã®17%ã§ãããªããagreementãafferctãemotinal suppportãªã©ã®socio-emotionalã¯33%ããããã®ã«å¯¾ãã¦*25ã
best answerãé¸ã¶ãã¨ã®ããä¸ã¤ã®ç¹å¾´ã¯ä»ã®å¤ãã®è¯ãåçãé¸ã°ããªãã¨ãããã¨ãå
è¡ãã¦è¡ã£ãå®é¨ã§ã¯ProgrammingãCancerãCelebrityãããããã100ã®è³ªåãã¨ã£ã¦æ¥ã¦ãåçã«æ¡ç¹ããããæ¬å½ã®best answerã¯ãããã«bestã§ãã£ãããbest answerã«é¸ãã è
ã¨ã¯éã£ã¦ããã2çªç®ã3çªç®ã«ããåçã ã£ãã
ã¤ã¾ãè¯ãåçã°ããããã¦ã¼ã¶ã¼ãè¦ã¤ãåºããã¨ãã§ããªããåçã®å¤ãã«ãã´ãªã¼ã§ã¯best answerã«é¸ã°ãã確çãå°ãããªã£ã¦ãã¾ãã
ããã§ãæè¡çãªãã¨ãäºå®ã«ã¤ãã¦ã®ã«ãã´ãªã¼ã§ã¯ä½ãentropyãperformanceã«é¢ä¿ããã¨expectããã
ããã証æããããã«entropyã2ã¤ç®ã®é層ãåãã¦è¨ç®ããã
ãã®çµææè¡çãªã«ãã´ãªã¼ã§ããComputers & Internetã¨Science & Mathã§ã¯ãããããªé¢ä¿ãè¦ãããã
å¼±ã¾ãã¯ãããadvice-ladenã®ã«ãã´ãªã¼ã§ããFamily & Relationshipsã§ãé¢ä¿ãè¦ãããã
Sportsã§ã¯é¢ä¿ã¯ãªãã£ãã
æå¾ã«ã«ãã´ãªã¼å
ã§ã®ã¦ã¼ã¶ã¼ã®åçã®å²åã¨best answerã«é¸ã°ããçããã¹ã¦ã®ã«ãã´ãªã¼ã§é¢ä¿ã¥ããã
æè¡çãªã«ãã´ãªã¼ã§ã¯focusã¯betterãªscoreã¨é¢ä¿ãã¦ãããåçã«ç¥èãå¿
è¦ãªã«ãã´ãªã¼ã§ãå¼±ã¾ãã¯ãããæãããªé¢ä¿ããã£ããdiscussionã«ãã´ãªã¼ã§ã¯å
¨ãé¢ä¿ããªãã£ãã
asker-repliier overlapãä½ããthread lengthãçãã«ãã´ãªã¼ãé«ãé¢ä¿ãæã£ã¦ããã
6. PREDICTING BEST ANSWERS
ããã¤ãã®æ¹æ³ã§ãã®åçãbest answerã«é¸ã°ãããã¨ã©ãããäºæ¸¬ãããã¨ãã§ããããã¹ããããconplementaryã§concurrentã«ã質åã¨åçã®qualityã«é¢ããç 究ãAgichiteinãã«ããè¡ããã¦ãã*26ã
ã©ã³ãã ã«best answerã¨ããã§ãªããã®ãbalanceããããã«åçãæ½åºããvery likelyã«bestã«é¸ã°ããããªåçãé¤å¤ããä¸ã§ãããã¤ãã®å¤æ°ã§ãã¸ã¹ãã£ãã¯å帰ãè¡ã£ãã
åçè
ã®å¤§åã¯åçæ°ãå°ãªãããã®ã§entropyã¨focusã®measuerã¯ä½¿ããªãã£ãã
äºæ¸¬ç²¾åº¦ãå¾ãããã«10åã®ã¯ãã¹ç¢ºèªãrandom guessesã§0.5ã®baselineã§è¡ã£ãã
Programmingã¨Marriageã¨Wrestlingã§è¡ã£ãã
åçã®é·ãã¨ãä»ã®åçã®å¤ãã大ããå½±é¿ããã
åçã®é·ãã ãã§62%ã®äºæ¸¬ç²¾åº¦ãå¾ãã質åè
ãããé·ãåçã好ããªãã¨ã示ãã¦ããã
ã¦ã¼ã¶ã¼ããã®ã«ãã´ãªã¼ã§åçãã¦ããæ°ã¨best answerã«é¸ã°ãã¦ããæ°ãè¯ãäºæ¸¬ãçã¿ãããã¯Programmingã§ä»ã®ã«ãã´ãªã¼ããé¡èã ã£ããä»ã®ã«ãã´ãªã¼ã§best answerã«é¸ã°ãã¦ããæ°ã¯é¢ä¿ãªãã£ãã
åç´ãªã¦ã¼ã¶ã¼ã®åçæ°ã¯ã»ãã®å
ãã«æ¹åãããããbest answerã®æ°ãèæ
®ããã¨negativeãªå½±é¿ãåã¼ãã
ãã®çµæã¯ä»¥åã«è¡ã£ãSun's Java Forumã§ã®çµæã¨éç«ã£ã対象ããªãã¦ãããJava Forumã§ã¯åçã®æ°ã¯expertise levelã¨å¼·ãé¢ä¿ãã¦ããã
ã¦ã¼ã¶ã¼ã®é¸ãã best answerã¨ãã®é åã®expertãé¸ãã åçã¨ã競äºãããã®ã¯é¢ç½ãããåçã®é »åº¦ã¨expertise levelã®é¢ä¿ãããã®ããJava Forumã®ãããªå°éçãªcommunityã¨éãYahoo Answerã®ãããªä¸è¬çãªcommunityã§ã¯expertise levelã«å¤§ããªéããããã®ããªã©ã調ã¹ãã®ãé¢ç½ããããããã¯future workã
7. CONCLUSIONS
ã¾ã
content propertyã¨ã«ãã´ãªã¼ãã¾ããã social network interactionãæ¯è¼ããthread lengthã¨overlapã«ããã«ãã´ãªã¼ãã¯ã©ã¹ã¿ãªã³ã°ã§ãããã¨ãfindããã
discussion topicãäºå®ã«åºã¥ãåçãæ±ãã¦ããªããããªtopicã§ã¯é·ãthreadã§activity levelã®åå¸ã¯å¹
åºããã¦ã¼ã¶ã¼ã¯è³ªåãåçãããå¾åããã£ãã
äºå®ãæ±ãã質åã®å¤ãã«ãã´ãªã¼ã§ã¯thread lengthã¯çããå
¸åçã«ã¦ã¼ã¶ã¼ã¯åãforumã§helperãaskerã«å¾¹ãã¦ããã
ãããã®ç°ãªãdynamicsã«differing interaction motifãcorrespondãããã¨ãfindããã
online forumã§ã®å
è¡ç 究ã¨åãããã«ãego-networkãdiscussion threadãæ¯é
ããå¾åã®ããYA categoryããquestion-answerå½¢å¼ã«ç¸ããã¦ããä¸ã§ãããç°¡åã«revealãããã¨ãfindããã
次ã«
é¢ä¿ã®ããã«ãã´ãªã¼ãidentifyãããããã«ãã´ãªã¼ã§åçããã¦ã¼ã¶ã¼ãä»ã®ã«ãã´ãªã¼ã§ãåçããããããããã¹ããã¨ã«ããã
質åã¨åçãå¥ã
ã«èããã¨ãé¢ç½ãé対称ãè¦ã¤ãã£ããfamiliar topicã«é¢ãã質åã«ã¯å¤ãã®ã¦ã¼ã¶ã¼ãåçãããå¤ã質åããã®ãã©ãã§ããspecializedã§technicalãªã«ãã´ãªã¼ã§åçãããããªã¦ã¼ã¶ã¼ã¯ä»ã®ã«ãã´ãªã¼ã§ã¯è³ªåãå°ãªãã
ã¦ã¼ã¶ã¼ãã©ãã ãã«ãã´ãªã¼ãã¾ããã§knowledgeãshareãã¦ãããã調ã¹ãå¤ãã®ã¦ã¼ã¶ã¼ã¯å¤ãã®éã£ãã«ãã´ãªã¼ã§åçãã¦ãããspecializedã§technicalãªã«ãã´ãªã¼ã§ã¯ãã®å¾åãå°ãªãã£ãããã®ãããªã«ãã´ãªã¼ã§ã¯ãã®ã«ãã´ãªã¼ã«focusãã¦ããã¦ã¼ã¶ã¼ã®æ¹ãbest answerã«é¸ã°ããããã£ãã
æå¾ã«
best answerãäºæ¸¬ãããã¨è©¦ã¿ãããã®è³ªåã¸ã®åçã®æ°ã¨åçè
ã®track recordã¨åæ§ã«åç´ã«åçã®é·ãããã£ã¨ãäºæ¸¬ã«ä½¿ãããã¦ã¼ã¶ã¼ã®best answerã®æ°ï¼expertizeã®potentialãªindicatorï¼ã¯å½¹ã«ç«ã£ãããããæãç·ã ã£ãã®ã¯technically focused Programming categoryã ã£ãã
future workã¨ãã¦YAã§shareããã¦ããexpertiseã®levelããã£ã¨èª¿ã¹ãããæ°ä¸»åããknowledge sharingã«ããYAã¯å¤§ããªåæ¥ãéæãããã¿ããªä½ãããã£ã¦ãããããã¦è§£æã«ããå¤ãã®äººãããã¤ãã®ãã¨ãç¥ã£ã¦ãã¦ãããYAã§shareã§ãããã¨ãç¥ã£ããããããã®å¹
åºããdepthãç ç²ã«ãã¦ãããã©ããã¯æªã unclearãæã
ãæ¥ã
åç´ãªè³ªåã®åçãå¾ã¦ããã®ã«å¯¾ãã¦ãtop levelã®expertãéã£ãincentive mechanismã«ããYAã«åå ãã¦ããã®ãã©ãããç¥ãããã
*1:T. Holloway, M. Bozicevic, and K. B¨orner. Analyzing and visualizing the semantic coverage of wikipedia and its authors: Research articles. Complexity, 12(3):30-40, 2007.
*2:S. Whittaker, L. Terveen, W. Hill, and L. Cherny. The dynamics of mass interaction. Proceedings of the 1998 ACM conference on Computer supported cooperative work, pages 257-264, 1998.
*3:K. Zhongbao and Z. Changshui. Reply networks on a bulletin board system. Phys. Rev. E, 67(3):036117, Mar 2003.
*4:D. Fisher, M. Smith, and H. Welser. You Are Who You Talk To: Detecting Roles in Usenet Newsgroups. In HICSSâ06, 2006.
*5:T. Turner, M. Smith, D. Fisher, and H. Welser. Picturing Usenet: Mapping Computer-Mediated Collective Action. Journal of Computer-Mediated Communication, 10(4), 2005.
*6:E. Wegner. Communities of Practice: Learning, Meaning, and Identity, 1998.
*7:J. Preece, B. Nonnecke, and D. Andrews. The top five reasons for lurking: improving community experiences for everyone. Computers in Human Behavior, 20(2):201-223, 2004.
*8:J. S. Donath. Identity and deception in the virtual community. Communities in Cyberspace, pages 29-59, 1999.
*9:H. T. Welser, E. Gleave, D. Fisher, and M. Smith. Visualizing the signatures of social roles in online discussion groups. Journal of Social Structure, 8(2), 2007.
*10:W. Sack. Conversation map: a content-based Usenet newsgroup browser. In IUIâ00, pages 233-240, 2000.
*11:J. Arguello, B. S. Butler, L. Joyce, R. Kraut, K. S. Ling, and X. Wang. Talk to me: foundations for successful individual-group interactions in online communities. In CHIâ06, pages 959-968, 2006.
*12:E. Joyce and R. Kraut. Predicting Continued Participation in Newsgroups. Journal of Computer-Mediated Communication, 11(3):723-747, 2006.
*13:K. Lakhani and E. von Hippel. How open source software works:âfreeâ user-to-user assistance. Research Policy, 32(6):923-943, 2003.
*14:B. Butler. Membership Size, Communication Activity, and Sustainability: A Resource-Based Model of Online Social Structures. Information Systems Research, 12(4):346-362, 2001.
*15:J. Zhang, M. Ackerman, and L. A. Adamic. Expertise networks in online communities: structure and algorithms. In WWWâ07, pages 221-230, 2007.
*16:J. Zhang, M. S. Ackerman, and L. A. Adamic. Communitynetsimulator: Using simulations to study online community networks. In C & Tâ07, 2007.
*17:Q. Su, D. Pavlov, J. Chow, and W. Baker. Internet-scale collection of human-reviewed data. In WWWâ07, pages 231-240, 2007.
*18:S. Kim, J. S. Oh, and S. Oh. Best-Answer Selection Criteria in a Social Q&A site from the User-Oriented Relevance Perspective. presented at ASIST, 2007.
*19:T. Turner, M. Smith, D. Fisher, and H. Welser. Picturing Usenet: Mapping Computer-Mediated Collective Action. Journal of Computer-Mediated Communication, 10(4), 2005.
*20:H. T. Welser, E. Gleave, D. Fisher, and M. Smith. Visualizing the signatures of social roles in online discussion groups. Journal of Social Structure, 8(2), 2007.
*21:R. Milo, S. Shen-Orr, S. Itzkovitz, N. Kashtan, D. Chklovskii, and U. Alon. Network Motifs: Simple Building Blocks of Complex Networks. Science, 298(5594):824–827, 2002.
*22:R. Milo, S. Itzkovitz, N. Kashtan, R. Levitt, S. Shen-Orr, I. Ayzenshtat, M. Sheffer, and U. Alon. Superfamilies of evolved and designed networks. Science, 303:1538–1542, 2004.
*23:S. Wernicke and F. Rasche. FANMOD: a tool for fast network motif detection. Bioinformatics, 22(9):1152–1153, 2006.
*24:J. Zhang, M. Ackerman, and L. A. Adamic. Expertise networks in online communities: structure and algorithms. In WWWâ07, pages 221–230, 2007.
*25:S. Wernicke and F. Rasche. FANMOD: a tool for fast network motif detection. Bioinformatics, 22(9):1152–1153, 2006.
*26:E. Agichtein, C. Castillo, D. Donato, A. Gionis, and G. Mishne. Finding High-Quality Content in Social Media. WDSMâ08, 2008.