hlalosang ka ho eketsehileng

Karolong ena e reretsoe ho sebelisoa e le ya referense eo e neng e e-na le ho bala e le phetoang.

  • Selelekela (Karolo 2.1)

mofuta ofe e mong oa ho boloka e sa akarelletsoa khaolong ena e ethnography. Etsoe eketsehileng mabapi le ethnography a libaka dijethale bona Boellstorff et al. (2012) , le bakeng sa eketsehileng mabapi le ethnography a tsoakiloeng libaka dijethale le 'meleng bona Lane (2016) .

  • Ya data Big (Karolo 2,2)

Ha u repurposing ya data, ho na le maqheka 'meli ba kelello e ka u thusang ho utloisisa mathata a ka etsahala hore u ka' na teana le. Taba ea pele, u ka leka ho nahana sete ea datha loketseng bakeng sa ho bothata ba hao le bapisa hore ho sete ea datha tseo u li sebelisa. Joang ba e tšoanang le kamoo a ba fapaneng? Haeba u ne ke sa bokella ya data ya hao u ithata, ho na le monyetla o moholoanyane oa ho ba phapang pakeng tsa seo u se batlang le seo u nang le sona. Empa, u lokela ho etsa qeto ya hore ho se tšoane tsena ke tse sa reng letho kapa e khōlō.

Ea bobeli, hopola hore motho e mong a bōpa le bokella ya data ya hao ka lebaka le itseng. U lokela ho leka ho utloisisa ho beha mabaka a bona. Ea mofuta ona e se mo khutlisetsang bocheng-boenjiniere ka u thusa ho hlokomela mathata a ka khona le biases a dintlha tsa hao repurposed.

Ha ho na betere nakong e leng lesoha o re tlhaloso ea "data ya e khōlō", empa litlhaloso tse ngata ho bonahala eka ho nahana ka tse 3 Vs: (mohlala, bophahamo ba modumo, tse fapa-fapaneng, 'me lebelo Japec et al. (2015) ). Ho e-na le ho tsepamisa kelello litšobotsi tsa boitsebiso, tlhaloso ea ka e bua haholo eketsehileng mabapi le hore na ke hobane'ng ya data ileng ba bōptjoa.

kenyelletsoa ha lintlha tsa ka ya data 'muso tsa tsamaiso tsa ka hare dihlopha tsa ya data khōlō ke a batla tsela e sa tloaelehang. Batho ba bang ba entse Tabeng ena, kenyeletsa Legewie (2015) , Connelly et al. (2016) , le Einav and Levin (2014) . Etsoe ho eketsehileng ka bohlokoa ba ya data 'muso tsa tsamaiso tsa tsa patlisiso ya, sheba Card et al. (2010) , Taskforce (2012) , le Grusky, Smeeding, and Snipp (2015) .

Bakeng sa pono ea ho etsa lipatlisiso tsa tsamaiso tsa ka hare tsamaiso mmuso dipalopalo, haholo-holo Bureau palo US, sheba Jarmin and O'Hara (2016) . Bakeng sa buka bolelele kalafo ea lipatlisiso tsa tsamaiso tsa direkoto tsa nakong Lipalo-palo Sweden, sheba Wallgren and Wallgren (2007) .

Khaolong eo, ke bokhutšoanyane bapisoa phuputso e entsoeng tsa moetlo tse kang General Social Survey (GSS) to ya setjhaba ya data mohloli o mecha ea phatlalatso ea kang Twitter. Bakeng sa ho phethahetseng 'me bapisoa ka hloko pakeng tsa liphuputso tsa moetlo le data tsoang sechabeng mecha ea litaba, sheba Schober et al. (2016) .

  • E Tloaelehileng litsobotsi ya data khōlō (Karolo 2.3)

Litsobotsi tsee 10 ya data khōlō ba 'nile ba hlalosoa ka litsela tse sa tšoaneng tse fapaneng ke mefuta e fapaneng ya bangodi fapaneng. Mongolo o susumetsoa ke monahano o reng ka ho ea pele litaba tsena li kenyeletsa: Lazer et al. (2009) , Groves (2011) , Howison, Wiggins, and Crowston (2011) , boyd and Crawford (2012) , Taylor (2013) , Mayer-Schönberger and Cukier (2013) , Golder and Macy (2014) , Ruths and Pfeffer (2014) , Tufekci (2014) , Sampson and Small (2015) , Lewis (2015) , Lazer (2015) , Horton and Tambe (2015) , Japec et al. (2015) , le Goldstone and Lupyan (2016) .

Ho pholletsa le khaolo ena, Ke 'nile ka sebelisitse lentsoe lena bonoa moroeroe dijethale, e leng ke nahana ke batlang ho se nke lehlakore. E nngwe gape ratoa haholo bakeng sa bonoa moroeroe dijethale e labnyomok dijethale (Golder and Macy 2014) , empa e le Hal Abelson, Ken Ledeen, 'me Harry Lewis (2008) supa, e loketseng ho tse thathamisitsoeng eketsehileng ho sebelisitsoe lentsoe le mohlomong dijethale menwana. Ha u theha labnyomok, o tseba se etsahalang 'me labnyomok ba hao ba ka kakaretso bo amahanngoa le uena ka ho toba. Hoa tšoana le ha se 'nete bakeng sa bonoa moroeroe oa hao dijethale. Ha e le hantle, u siea mesaletsa eohle ka nako eohle eo u nang le tsebo e fokolang haholo. 'Me, le hoja mesaletsa eohle tsena ha ba na lebitso la hao ho bona, hangata ba ka ho amahanngoa khutlela ho uena. Ka mantsoe a mang, ba ka ho fetisisa jwalekaha menwana: bonahaleng le batho ka bomong le khethollang.

Big

Bakeng sa na ke hobane'ng ha datasets khōlō, li fetolela ua ngola litlhahlobo dipalopalo fana ka mathata, bona Lin, Lucas, and Shmueli (2013) le ho McFarland and McFarland (2015) . ditaba tsena lokela lebisa bafuputsi ho tsepamisa kelello boleloang bo sebetsang ho e-na le moelelo oa dipalopalo.

Kamehla-on

Ha le tšohla kamehla-on boitsebiso, ho ke ke habohlokoa ho nahana ka hore na u bapisa le tobileng batho tšoanang ha nako e ntse kapa hore na o bapisa tse ling sehlopha se fetola ea sechaba sa ka; sheba mohlala, Diaz et al. (2016) .

Non-arabelang

A buka khale pele ho mehato non-arabelang ke Webb et al. (1966) . Mehlala bo hlahang bukeng pele ho letsatsi la ho le lilemo li dijethale, empa ba ntse khantša. Bakeng sa mehlala ea batho ba fetola boitšoaro ba bona ka lebaka la ho ba teng ha boima leihlo, sheba Penney (2016) le ho Brayne (2014) .

sa fellang

Bakeng sa tse tlalehiloeng khokahano, sheba Dunn (1946) le ho Fellegi and Sunter (1969) (libaka tse tummeng historing) le Larsen and Winkler (2014) (eo kajeno). Similar atamela le tsona li qapa ho tsa saense khomphuteng ya tlas'a mabitso a kang ya data deduplication, mohlala boitsebahatso, lebitso la tshwanelelanang, etsisa bonoa, le etsisa rekoto phumano (Elmagarmid, Ipeirotis, and Verykios 2007) . Ho boetse ho na boinotši ba bolokoe le atamela hore a tlalehe khokahano e leng u se ke ua hloka phetiso ea batho ka bomong le khethollang fumane boitsebiso bo (Schnell 2013) . Facebook boetse e ntshetswa pele e tswela pele ho amahanya direkoto tsa bona ba le mekhoa vouta; sena se ne se etsoa ho lekola e teko eo ke tla u bolella hoo e ka bang ka Khaolo ea 4 (Bond et al. 2012; Jones et al. 2013) .

Etsoe eketsehileng mabapi le ho betla tsela tumello, sheba Shadish, Cook, and Campbell (2001) , Khaolong ea 3.

atameleha

Bakeng sa ka AOL fuputso kutu debacle, sheba Ohm (2010) . Ke fana ka keletso e mabapi le thekgago le dikhamphani le mebuso Khaolong 4 ha ke hlalosa liteko. Maloa ba bangodi ba boletse dingongoreho mabapi lipatlisiso tseo relies ya data atameleha, sheba Huberman (2012) le ho boyd and Crawford (2012) .

Tsela e 'ngoe molemo bakeng sa bafuputsi univesithi ho fumana phihlello ya data ke ho sebetsa k'hamphani e le boleng bo intern kapa ho etela mofuputsi. Ho phaella ho nolofalletsa phihlello ya data, tshebetso ena e tla thusa mofuputsi ea ithuta ho eketsehileng mabapi le kamoo boitsebiso ileng ba bōptjoa, e leng sa bohlokoa ho hlahlojoe.

Non-moemeli

Non-representativeness ke bothata bo boholo bakeng sa bafuputsi le mebuso ba lakatsang ho etsa lipolelo tse mabapi le e baahi kaofela. Sena ke ka tlaase ho moo la ho ameha ka lik'hamphani tse ba ka tloaelo ho tsepamisa maikutlo basebedisi ba bona. Etsoe eketsehileng mabapi le kamoo Lipalo-palo Netherlands o nahana ka taba ea ho bao e seng representativeness ya data likhoebo tse khōlō, sheba Buelens et al. (2014) .

Khaolong ea 3, ke tla hlalosa mehlala le phopholetso ka ho qaqileng lintho tse ngata e khōloanyane. Esita le haeba ya data tse bao e seng moemeli tlas'a maemo a itseng, ba ka weighted ho hlahisa likhakanyo tsa molemo.

hoholeha

System hoholeha thata haholo ho bona ho tloha ka ntle. Leha ho le joalo, e MovieLens projeke (tšohloa haholoanyane Khaolong ea 4) 'nile matha ka lilemo tse fetang 15 ke e le sehlopha se etsa lipatlisiso akatemi. Ka lebaka leo, ba tlaleha le ba ile ba kopanela tlhahisoleseding e mabapi le tsela eo tsamaiso e bile teng ka nako e telele 'me kamoo sena ka' na tshusumetso Analysis (Harper and Konstan 2015) .

Litsebi tse ngata tsa ba tsepame hoholeha a Twitter: Liu, Kliman-Silver, and Mislove (2014) le ho Tufekci (2014) .

Algorithmically na mobulungano

Ke qala ho utloa lentsoe "algorithmically na mobulungano" sebelisoang ke Jon Kleinberg puo. Khopolo se seholo ka morao performativity ke hore lilemong tse ka bang saense ea bitsoang likhopolo tsa phedisano le "dienjini eseng lik'hamera" (Mackenzie 2008) . Ke hore, ha e le hantle ama lefatše ho e-na feela hapa.

ditshila

'Muso, mekhatlo e dipalopalo bitsa ya data ho hloekisa, ho dipalopalo ya data tseleganya. De Waal, Puts, and Daas (2014) e hlalosa botegeniki jwa dipalopalo ya data tseleganya ntshetswa pele bakeng sa data phuputso e entsoeng le ho hlahloba eo isa tekanyong ba sebetsang ho mehloling khōlō ya data, le Puts, Daas, and Waal (2015) hlahisang tse ling tsa likhopolo tse neng e le 'ngoe bakeng sa bamameli ba eketsehileng kakaretso.

Etsoe ka mehlala e meng ea lithuto tsa tsepamisa maikutlo spam a Twitter, Clark et al. (2016) le ho Chu et al. (2012) . Qetellong, Subrahmanian et al. (2016) e hlalosa liphello tsa DARPA Twitter Bot Bothata.

mamelang le e ikemiselitseng

Ohm (2015) lekodisisa etsa lipatlisiso pejana ka khopolo ea ho fumana boitsebiso bo mamelang le e ikemiselitseng le e fana ka selotlolo Ntlha e teko. The lintlha tse 'nè o ile a proposes ke: kgonego ya ntša kotsi; kgonego ya ntša kotsi; boteng ba le kamano e lekunutu; le hore na kotsi bonahatsa dingongoreho majoritarian.

  • Bala lintho (Karolo 2.4.1)

Thuto Farber ea taxis New York e ne e thehiloe ho ithuta ka eena pejana ke Camerer et al. (1997) hore sebelisoa tharo disampole fapaneng bebofalletsa hatisitsoeng pampiring tsamaee leeto la lik'hilomithara lishiti tsa pampiri mefuta e sebelisoa ke bakhanni hore a tlalehe tsamaee leeto la lik'hilomithara qale nako, qetellong nako, 'me ba tsoella. Thutong ena pejana fumane hore bakhanni ba bonahala e le shebiloeng earners: ba ile ba sebetsa ka tlaase ho moo ka matsatsi a moo moputso bona e ne e phahameng.

Kossinets and Watts (2009) e ile tsepame tšimoloho ea homophily a websaeteng. Sheba Wimmer and Lewis (2010) ea ho atamela e fapaneng bothateng bona tšoanang e leng sebedisa data ho tswa ho Facebook.

Mosebetsing morago ga moo e, Morena le 'basebetsi-'moho ba ka ho eketsehileng tsa utolola Inthaneteng li thibela Chaena (King, Pan, and Roberts 2014; King, Pan, and Roberts 2016) . Ea ho atamela e amanang le lefu la ho lekanya Inthaneteng li thibela Chaena, sheba Bamman, O'Connor, and Smith (2012) . Etsoe eketsehileng mabapi le mekhoa ea dipalopalo ho tšoane le le le sebelisitsoeng ho King, Pan, and Roberts (2013) ho hakanya ea maikutlo ea lipalong tse limilione tse 11, sheba Hopkins and King (2010) . Bakeng sa pele ho ithuta a laola, sheba James et al. (2013) (ka tlase botekgeniki) le Hastie, Tibshirani, and Friedman (2009) (eketsehileng botekgeniki).

  • Bonelang pele (Karolo 2.4.2)

Bonelang ke karolo e khōlō ea data ho liindasteri tsa saense ea bitsoang (Mayer-Schönberger and Cukier 2013; Provost and Fawcett 2013) . Mofuta o mong oa mesebetsi bonelang tse atisang ho etsoa ke bafuputsi ba tsoang sechabeng ke babapatsi bonelang, ho etsa mohlala Raftery et al. (2012) .

Google Ntaramane Trends e ne e se projeke ya pele ya ho sebedisa ya data ho phenyekolla ho nowcast ata ntaramane. Ha e le hantle, bafuputsi ba United States (Polgreen et al. 2008; Ginsberg et al. 2009) le Sweden (Hulth, Rydevik, and Linde 2009) ba fumane hore dipehelo tse itseng fuputso (mohlala, "ntaramane") a bolela esale pele naha ea phatlalatsa leihlo bophelo bo botle ya data pele e ile ea lokolloa. O ile a mangata, merero e meng e mengata 'nile ba leka ho sebelisa ya data dijethale moneketsela bakeng phumano leihlo maloetse, bona Althouse et al. (2015) bakeng sa tekolo botjha e.

Ho phaella ho sebelisa ya data dijethale moneketsela ho noha diphetho tsa bophelo bo botle, ho na le e boetse e 'nile ea e e le mosebetsi sebelisa ya data Twitter ho noha diphetho dikgetho; etsoe recensioni bona Gayo-Avello (2011) , Gayo-Avello (2013) , Jungherr (2015) (CH 7.), 'me Huberty (2015) .

U sebelisa ya data ho phenyekolla ho boleloa esale pele ka ata ntaramane le ho sebelisa ya data Twitter ho noha likhethong ke mehlala bobeli ba sebelisa mofuta o itseng oa dijethale moneketsela ho noha ea mofuta o itseng ketsahalo lefatšeng. Ho na le nomoro ya e ngata haholo ea lithuto tsa hore na sebopeho sena kakaretso. Lethathamo 2,5 akarelletsa mehlala e seng mekae e meng.

Lethathamo 2,5: Leeme lethathamo la lithuto tsa sebelisa lilemong tse ka bang moneketsela dijethale ho noha ba bang ketsahalo eo.
Digital moneketsela sephetho sa ho qotse
Twitter Box matlotlong ofisi ea lifilimi ka US Asur and Huberman (2010)
Search notlelela Thekiso ea lifilimi, 'mino, libuka, lipapali tsa video li ka US Goel et al. (2010)
Twitter Dow Jones Industrial Karolelano (US marakeng) Bollen, Mao, and Zeng (2011)
  • Approximating liteko (Karolo 2.4.3)

Koranta ea PES Science Lipolotiki ne neheletsano ya data e khōlō, e bakang diplomat, 'me khopolo hakaalo,' me Clark and Golder (2015) akaretsa mong le e mong ka monehelo. Koranta ea ditshebetsong tsa Academy of Sciences National ea United States of America e ne e neheletsano mabapi le diplomat e bakang le data khōlō, 'me Shiffrin (2016) akaretsa mong le e mong ka monehelo.

Ho latela liteko tsa tlhaho, Dunning (2012) e fana ka leseli le babatsehang bukeng bolelele kalafo. Etsoe eketsehileng mabapi le ho sebelisa Vietnam moralo lothari e le teko ea tlhaho, sheba Berinsky and Chatfield (2015) . Etsoe le atamela mochine ithuta hore leka ho ka tsela e iketsang hlokomela liteko tsa tlhaho hare ho mehloling khōlō ya data, sheba Jensen et al. (2008) le ho Sharma, Hofman, and Watts (2015) .

Ho latela le tshwanang, bakeng sa tekolo botjha le tšepo, sheba Stuart (2010) , le bakeng sa ho lekolwa botjha-nyatsa bona Sekhon (2009) . Etsoe eketsehileng mabapi le matching joalokaha e mosa ea faolang, sheba Ho et al. (2007) . Etsoe libuka tse fanang ka mekhoa ea phekolo e babatsehang ea tshwanelelanang, sheba Rosenbaum (2002) , Rosenbaum (2009) , Morgan and Winship (2014) , le Imbens and Rubin (2015) .