[ , ] Algorithmic confounding yog ib qho teeb meem nrog Google Flu Trends. Nyeem daim ntawv los ntawm Lazer et al. (2014) , thiab sau luv luv, meej email rau tus kws kho tsheb ntawm Google piav qhov teeb meem thiab muab tswv yim ntawm kev txhim kho nws.
[ ] Bollen, Mao, and Zeng (2011) claims tias cov ntaub ntawv los ntawm Twitter siv tau los kwv yees qhov kev ua lag luam Tshuag. Qhov kev tshawb nrhiav no tau tsim rau kev tsim ntawm lub pob nyiaj-Derwent Capital Markets-los nqis peev rau hauv kev ua lag luam Tshuag raws li cov ntaub ntawv sau los ntawm Twitter (Jordan 2010) . Cov pov thawj twg yog qhov koj xav pom ua ntej tso koj cov nyiaj rau hauv cov nyiaj ntawd?
[ ] Thaum qee tus neeg pabcuam pejxeem saib xyuas kev nojqab haushuv xav txog e-luam yeeb pabcuam kev haus luamyeeb, lwm tus tau ceeb toom txog kev pheej hmoo, xws li cov qib siab nicotine. Xav txog tias ib tug neeg tshawb nrhiav txiav txim siab los kawm txog pej xeem lub tswv yim mus rau e-luam yeeb los ntawm kev sau cov e-luam yeeb uas muaj feem xyuam nrog Twitter thiab ua cov kev xav sib txawv.
[ ] Nyob rau lub Kaum Ib Hlis 2009, Twitter tau hloov cov lus nug hauv kem thawv tawm ntawm "Koj ua dab tsi?" Rau "Qhov teeb meem tshwm sim?" (Https://blog.twitter.com/2009/whats-happening).
[ ] "Retweets" feem ntau siv los ntsuas kev cuam tshuam thiab kev sib kis ntawm lub peev xwm ntawm Twitter. Pib, cov neeg siv yuav tsum luam thiab twv lawv cov twv lawv nyiam, lossis sau tus thawj tus neeg sau nrog nws tus kheej, thiab ntaus hom ntawv "RT" ua ntej tweet los qhia tias nws yog ib qho kev sib tw. Tom qab ntawd, xyoo 2009, Twitter tau ntxiv "khawm" retweet. Nyob rau hauv Lub Rau Hli 2016, Twitter tau ua rau cov neeg siv mus retweet lawv cov tweets (https://twitter.com/twitter/status/742749353689780224). Koj puas xav tias cov kev hloov no yuav cuam tshuam rau koj siv "retweets" hauv koj qhov kev tshawb fawb? Vim li cas los yog vim li cas?
[ , , , ] Hauv kev sib tham dav dav, Michel thiab cov npoj yaig (2011) tau soj ntsuam cov ntsiab lus ntawm ntau tshaj tsib lab tus lej hauv cov ntawv hauv kev sim kom pom cov kev coj noj coj ua mus ntev. Cov ntaub ntawv uas lawv siv tau tam sim no tau tso tawm raws li Google NGrams dataset, thiab yog li ntawd peb thiaj li siv tau cov ntaub ntawv los xyuas dua thiab txuas ntxiv ntawm lawv cov hauj lwm.
Nyob rau hauv ib qho ntawm ntau cov qhabnias hauv daim ntawv, Michel thiab cov npoj yaig sib cav hais tias peb tsis nco qab sai thiab sai dua. Rau xyoo xyov, hais tias "1883," lawv suav qhov kev faib ua feem ntawm 1-grams luam tawm txhua xyoo ntawm xyoo 1875 thiab 1975 uas yog "1883". Lawv tau txiav txim siab tias qhov kev faib ua feem no yog ib qho kev ntsuas ntawm cov paj laum hauv cov xwm txheej uas tshwm sim rau xyoo ntawd. Nyob rau hauv lawv daim duab 3a, lawv tau npaj cov kev siv cuab yeej siv rau peb lub xyoo: 1883, 1910, thiab 1950. Cov peb xyoos no sib txawv xws li: siv me me ua ntej xyoo ntawd, tom qab ntawd ntev, ces poob. Tom ntej no, yuav kom muaj ntau npaum li cas ntawm txhua xyoo, Michel thiab cov npoj yaig suav txhua xyoo ntawm xyoo 1875 thiab 1975. Hauv lawv daim duab 3a (inset), lawv pom tias lub neej ib nrab ntawm txhua xyoo yog zuj zus, thiab lawv tau sib cav tias qhov no txhais tau hais tias peb tsis nco qab txog yav dhau los sai dua thiab sai dua. Lawv siv Version 1 ntawm cov lus Askiv corpus, tab sis tom qab Google tau tso tawm ib thib ob version ntawm lub corpus. Thov nyeem tag nrho cov lus nug ntawm koj ua ntej koj pib coding.
Yam ntxim saib ntxim ua no yuav muab rau koj xyaum sau cov cai siv tau, kev txhais cov ntsiab lus, thiab cov ntaub ntawv sib txuas lus (xws li ua haujlwm nrog cov ntawv ua tau txawv thiab tuav cov ntaub ntawv tsis muaj tseeb). Qhov kev ua no kuj yuav pab kom koj sawv thiab khiav nrog ib tus nplua nuj thiab nyiam dataset.
Tau txais cov ntaub ntawv raw ntawm Google phau ntawv NGram Viewer lub website. Hauv particular, koj yuav tsum siv version 2 ntawm lus Askiv corpus, uas tau tso tawm rau Lub Xya hli ntuj 1, 2012. Uncompressed, cov ntaub ntawv no 1.4GB.
Txaus siab rau qhov tseem ceeb ntawm daim duab 3a ntawm Michel et al. (2011) . Rov qab ua dua daim duab no, koj yuav tsum muaj ob cov ntaub ntawv: qhov koj tau downloaded hauv ib feem (a) thiab cov "tag nrho suav" cov ntaub ntawv, uas koj tuaj yeem siv hloov cov nyaj nyaj rau kev faib tawm. Nco ntsoov tias tag nrho suav cov ntaub ntawv muaj ib tus qauv uas ua rau nws nyuaj rau nyeem ntawv hauv. Puas version 2 ntawm cov ntaub ntawv NGram ua cov ntawv zoo sib xws rau cov neeg muaj npe hauv Michel et al. (2011) , uas yog raws li version 1 cov ntaub ntawv?
Tam sim no kos koj lub teeb graph lub teeb uas tsim los ntawm NGram Viewer.
Recreate daim duab 3a (daim duab loj), tab sis hloov cov \(y\) -axis ua tus zauv hais suav (tsis tus nqi ntawm cov lus sau).
Puas yog qhov txawv ntawm (b) thiab (d) coj koj rov mus soj ntsuam ib qho ntawm Michel li al. (2011). Vim li cas los yog vim li cas?
Tam sim no, siv qhov kev faib ua feem ntawm cov lus hais, muab tsab ntawv tso tawm ntawm daim duab 3a. Ntawd yog, rau txhua xyoo ntawm xyoo 1875 thiab 1975, xam lub sijhawm ib nrab ntawm xyoo ntawd. Ib nrab ntawm lub neej yog txhais tias yog cov xyoo dhau los ua ntej qhov feem pua ntawm cov lus hais mus txog ib nrab ntawm cov nqi siab tshaj. Nco ntsoov tias Michel et al. (2011) ua ib yam dab tsi nyuab dua rau kwv yees li ib nrab ntawm lub neej - saib Ntu III.6 ntawm Cov Lus Qhia Hauv Kev Pabcuam Online-tab sis lawv thov tias ob qho kev ua tau zoo ib yam. Puas version 2 ntawm NGram cov ntaub ntawv ua tau zoo ib yam li cov neeg nyob hauv Michel et al. (2011) , uas yog raws li version 1 cov ntaub ntawv? (Hint: Tsis txhob xav tsis thoob yog tias nws tsis haum.)
Puas muaj ib lub xyoo twg uas tsis tau muaj hnub nyoog xws li xyoo uas tau hnov qab tshwj xeeb sai los sis qeeb qeeb? Qhia me ntsis txog qhov laj thawj ntawm qhov ua tau zoo thiab piav qhia seb koj tau txheeb xyuas qhov tseeb.
Tam sim no replicate no tshwm sim rau version 2 ntawm NGrams cov ntaub ntawv nyob rau hauv Suav, Fabkis, German, Hebrew, Italian, Lavxias thiab Mev.
Sib piv hauv txhua hom lus, puas muaj txhua xyoo uas tau tawm tsam, xws li lub xyoo uas tau hnov qab tshwj xeeb los yog qeeb qeeb? Qhia me ntsis txog cov laj thawj ntawm cov qauv ntawd.
[ , , , ] Penney (2016) tau tshawb xyuas seb puas muaj kev tshaj tawm txog NSA / PRISM surveillance (ie, Snowden cov tshwm sim) nyob rau lub Rau Hli Ntuj 2013 tau txuam nrog kev kub nyhiab thiab cia li txo cov tsheb mus rau cov xov xwm hauv cov ncauj lus uas nce kev txhawj xeeb txog kev ceev ntiag tug. Yog li ntawd, qhov kev hloov hauv tus cwj pwm yuav zoo ib yam li cov nyhuv uas tshwm sim los ntawm huab hwm coj. Txoj kev ntawm Penney (2016) muaj qee zaus hu ua lub sij hawm tsim ntu xov xwm , thiab nws muaj feem xyuam nrog cov kev qhia uas tau piav qhia hauv seem 2.4.3.
Yuav xaiv lub ntsiab lus tseem ceeb, Penney raug xa mus rau cov npe uas siv los ntawm Lub Tsev Haujlwm Saib Xyuas Kev Ruaj Ntseg Hauv Tebchaws Asmeskas rau kev nrhiav thiab kev saib xyuas kev sib raug zoo. DHS sau npe qee cov ntsiab lus tshawb nrhiav mus rau ntau qhov teeb meem, xws li, "Kev Nyab Xeeb Kev Nyab Xeeb," "Kev Nyab Xeeb Kev Nyab Xeeb," thiab "Kev Phem Siab" Rau pawg neeg tshawb nrhiav, Penney siv 48 nqe lus hais txog "Kev Phem Siab" (saib cov ntawv appendix 8) ). Nws ces tag nrho Wikipedia Wikipedia pom suav txhua lub hli rau kev coj 48 Wikipedia cov ntawv dhau ib lub sij hawm 32-lub hlis, txij thaum pib ntawm lub Ib Hlis 2012 txog rau thaum xaus ntawm lub yim hli ntuj xyoo 2014. Txhawb zog nws qhov kev sib cav, nws kuj tau tsim ntau pawg sib txawv los ntawm kev nrhiav tsab xov xwm views rau lwm cov ncauj lus.
Tam sim no, koj yuav rov luam thiab txuas ntxiv Penney (2016) . Tag nrho cov ntaub ntawv xov raw uas koj xav tau rau qhov kev ua no yog muaj nyob ntawm Wikipedia. Los yog koj tuaj yeem tau txais los ntawm R-pob wikipediatrend (Meissner and R Core Team 2016) . Thaum koj sau koj cov lus teb, thov nco tias cov ntaub ntawv twg koj tau siv. (Nco ntsoov tias kev ua haujlwm tib yam no tshwm nyob rau hauv tshooj 6.) Qhov kev ua no yuav qhia koj xyaum hauv cov ntaub ntawv sib cav thiab xav txog tej yam kev sim hauv cov ntaub ntawv loj. Nws tseem yuav tau txais koj thiab khiav nrog cov ntaub ntawv uas muaj peev xwm tshaj plaws rau yav tom ntej tej yaam num.
[ ] Efrati (2016) qhia, raws li cov ntaub ntawv npog cia, tias "tag nrho cov sib koom" hauv Facebook tau txo los ntawm 5.5% xyoo dhau xyoo thaum lub sij hawm "kev tshaj xov xwm qub" tau nqis tshaj 21% xyoo dhau xyoo. Cov kev poob qis no tshwj xeeb yog mob nrog cov neeg siv Facebook hauv hnub nyoog 30 xyoos. Daim ntawv qhia tau hais tias qhov poob rau ob yam. Ib qho yog txoj kev loj hlob ntawm cov "phooj ywg" neeg muaj nyob hauv Facebook. Qhov thib ob yog tias qee qhov kev koom ua ke tau hloov rau kev tshaj tawm thiab cov neeg sib tw xws li Snapchat. Daim ntawv tshaj tawm kuj tau qhia txog ntau lub tswv yim Facebook tau sim ua kom muaj kev sib koom ua ke, xws li Xov xwm Feed algorithm tweaks uas yog cov thawj cov ntawv tseem ceeb dua qub, thiab raws li periodic nco ntawm thawj cov posts nrog "Nyob Rau Hauv Hnub" feature. Dab tsi ua rau, yog tias muaj, cov kev tshawb nrhiav no muaj rau cov kws tshawb fawb uas xav siv Facebook ua ib qhov chaw muab cov ntaub ntawv?
[ ] Qhov txawv ntawm ib tug kws lij choj thiab keeb kwm dab tsi? Raws li Goldthorpe (1991) , qhov sib txawv loj yog tswj cov ntaub ntawv khaws tseg. Keeb kwm cov neeg raug yuam kom siv cov kev pab cuam, qhov twg sociologists tuaj yeem hloov lawv cov ntaub ntawv sau rau cov laj thawj. Nyeem Goldthorpe (1991) . Qhov kev sib txawv ntawm kev ua neej thiab keeb kwm muaj feem xyuam rau lub tswv yim ntawm custommades thiab cov lus npaj tseg li cas?
[ ] Qhov no ua rau dhau los ntawm kev quesiton. Goldthorpe (1991) tau rub tawm ntau cov lus teb tseem ceeb, nrog rau ib qho ntawm Nicky Hart (1994) uas tau sib tw Goldthorpe txoj kev rau siab rau cov ntaub ntawv. Kom paub meej txog cov kev txwv ntawm cov qauv ntaub ntawv, Hart qhia txog Affluent Worker Project, kev tshawb fawb loj los ntsuas qhov kev sib raug zoo ntawm cov chav kawm ntawv thiab xaiv tsa uas tau ua los ntawm Goldthorpe thiab cov neeg ua haujlwm hauv lub xyoo 1960. Raws li ib tus xav tau los ntawm ib tus kws tshawb fawb uas tau tsim cov ntaub ntawv los nrhiav tau cov ntaub ntawv, Affluent Worker Project tau khaws cov ntaub ntawv uas tau teem tseg rau kev xav txog qhov kev xav txog yav tom ntej ntawm kev kawm nyob rau hauv ib lub sijhawm ntawm cov qauv kev noj haus. Tab sis npog, Goldthorpe thiab cov npoj yaig ua cas "tsis nco qab" mus sau cov lus qhia txog kev coj tus cwj pwm ntawm cov poj niam. Ntawm no yog li cas Nicky Hart (1994) tau qhia qhov tseem rov qab:
"... nws [yog] tsis yooj yim kom tsis txhob txiav txim siab tias cov poj niam raug rho tawm vim hais tias qhov no 'tailor made' dataset tau tawm los ntawm ib qho kev ua tau zoo uas tsis yog poj niam txoj kev. Tsav los ntawm lub tswvyim theoretical thiab kev ua haujlwm raws li txiv neej ua ntej ..., Goldthorpe thiab nws cov neeg ua haujlwm ua ke tsim ib co kev ua pov thawj pom uas tau noj thiab nurtured lawv tus kheej theoretical cov kev xav es tsis txhob raug lawv mus rau ib qho kev xeem ntawm kev txaus. "
Hart txuas ntxiv:
"Qhov kev tshawb nrhiav ntawm tus Affubent Worker Project qhia peb txog qhov tseem ceeb ntawm cov qauv kev paub ntawm cov qauv ntawm lub xyoo pua es tshaj qhov lawv qhia txog cov txheej txheem ntawm kev ua haujlwm, kev lag luam thiab kev siv hauv lub neej."
Koj tuaj yeem xav txog lwm cov kev ua pivtxwv uas cov ntaub ntawv tailor-ua muaj qhov kev sib tw ntawm cov ntaub ntawv collector ua rau nws? Qhov no puas piv rau algorithmic confounding li cas? Dab tsi ua rau qhov no muaj rau thaum cov neeg tshawb xyuas yuav tsum siv cov lus npaj tseg thiab thaum twg lawv yuav tsum siv cov custommades?
[ ] Hauv tshooj lus no, kuv tau hloov cov ntaub ntawv sau los ntawm cov neeg tshawb xyuas rau cov neeg tshawb xyuas txog kev tswj cov ntaub ntawv tsim los ntawm cov tuam txhab thiab cov tsoom fwv. Qee cov tib neeg hu rau cov ntaub ntawv hais txog "administrative data", uas lawv txawv txav nrog "tsim cov ntaub ntawv." Nws muaj tseeb tias cov ntaub ntawv tuav ntaub ntawv raug tshawb pom los ntawm cov neeg tshawb xyuas, tab sis lawv kuj tsim muaj. Piv txwv li, niaj hnub tech companies ua hauj lwm nyuab heev rau sau thiab piav txog lawv cov ntaub ntawv. Li no, cov ntaub ntawv cov ntaub ntawv no yog tsim thiab tsim, nws tsuas yog nyob ntawm koj qhov kev xav (duab 2.12).
Muab ib qho piv txwv ntawm cov ntaub ntawv qhov chaw uas pom nws ob qho tib si raws li tau pom thiab tsim muaj kev pab yog siv cov ntaub ntawv no los tshawb nrhiav.
[ ] Hauv kev xav, Christian Sandvig thiab Eszter Hargittai (2015) tau tshawb cov kev tshawb fawb los ua ob pawg ntsej muag nyob ntawm seb cov kab ke digital yog "ntsuas" lossis "kwv kawm ntawm kev kawm." Ib qho piv txwv ntawm thawj hom-qhov twg ib qho cuab yeej-yog qhov kev tshawb fawb los ntawm Bengtsson thiab cov npoj yaig (2011) siv cov xov tooj ntawm tes siv xov tooj los mus khiav cov kev tsiv teb tsaws tom qab av qeeg hauv Hiati xyoo 2010. Ib qho piv txwv ntawm ob hom-qhov chaw uas yog qhov khoom ntawm kev kawm-yog tshawb fawb los ntawm Jensen (2007) yuav ua li cas qhia txog cov xov tooj ntawm tes thoob plaws hauv Kerala, Is Nrias teb ua rau kev ua haujlwm ntawm kev ua lag luam rau cov ntses. Kuv pom tias qhov kev cais no pab tau vim tias nws qhia meej tias cov kev tshawb fawb los ntawm cov ntaub ntawv siv tau cov peev txheej muaj peev xwm txawm tias lawv siv tib hom ntaub ntawv. Yuav kom paub ntxiv txog qhov kev sib txawv ntawm no, piav txog plaub yam kev tshawb kawm uas koj tau pom: ob qho uas siv qhov system ua ib qho cuab yeej thiab ob qho uas siv lub digital system ua ib qho chaw kawm. Koj tuaj yeem siv cov qauv ntawm tshooj no yog koj xav tau.