imisebenzi

• ubunzima bobunzima: kulula , ephakathi , nzima , kakhulu kakhulu
• idinga izibalo ( ) U
• kufuna i coding ( ) U
• ukuqokelela idatha ( ) U
• zam intandokazi ( ) U

1. [U , U ] Ukuphazamiseka kwe-Algorithmic kwakuyinkinga ngeThans Trends zeGoogle. Funda iphepha Lazer et al. (2014) , kwaye ubhale i-imeyile emfutshane, ecacileyo kwinjineli kwi-Google ichaza ingxaki kwaye inike ingcamango yokuyilungisa.

2. [U ] Bollen, Mao, and Zeng (2011) bathi i-data evela kwi-Twitter ingasetyenziselwa ukuqikelela ukuthengisa imasheya. Oku kufumaniswe kukhokelela ekudalweni kwesikhwama se-Hedge Fund-Derwent Capital Markets-ukutyalomali kwimarike yemasheya esekelwe kwidatha eqokelelwe kwi-Twitter (Jordan 2010) . Yiyiphi ubungqina ofuna ukuyibona ngaphambi kokubeka imali yakho kwingxowa-mali?

3. [U ] Nangona abanye abasemagunyeni bempilo yoluntu bajonga i-e-cigarettes uncedo olusebenzayo lokuyeka ukutshaya, abanye bayalumkisa malunga neengozi ezinobungozi, njengamazinga aphakamileyo e-nicotine. Khawucinge ukuba umphandi unquma ukufundisisa uluntu ngokubhekiselele kwi-e-cigarettes ngokuqokelela izithuba ze-Twitter ezinxulumene ne-cigarettes nokuqhuba uhlalutyo lwamagama.

   1. Ziziphi ezi zintathu ezinokubangela ukhathazeka kakhulu kule sifundo?
   2. Clark et al. (2016) yaqhubela isifundo esinjalo nje. Okokuqala, baqokelela iifayili eziyi-850 000 ezisebenzisa i-e-cigaret-keywords ukusuka kwiJanuwari 2012 ukuya kuDisemba 2014. Xa behlolisiswa, bafumanisa ukuba ezininzi zezi tweets zazenzekelayo (okt, zingakhange ziveliswe ngabantu) kwaye ezininzi zale tweets zizenzekelayo i ntengiso. Baye baqulunqa i-algorithm yoluntu yokuhlukanisa iifayile ezizenzekelayo kwii-tweets eziphilayo. Ukusebenzisa lo mntu ukubona i-algorithm bafumene ukuba i-80% yee-tweets zazenzekelayo. Ngaba oku kufumanisa impendulo yakho kwingxenye (a)?
   3. Xa befanisa ukuvakalelwa kwii-tweets ze-organic kunye nezizenzekelayo, bafumanisa ukuba ii-tweets ezizenzekelayo zithandeka ngaphezu kweetweets ze-tweets (6.17 ngokumalunga no-5.84). Ngaba oku kufumana utshintsho impendulo yakho (b)?

4. [U ] NgoNovemba 2009, i-Twitter yatshintsha umbuzo kwibhokisi le-Twitter elivela "Yintoni oyenzayo?" Ukuba "Yintoni eyenzekayo?" (Https://blog.twitter.com/2009/whats-happening).

   1. Ucinga ukuba utshintsho lwezithuba luya kuthintela bani na i-tweets kunye / okanye yintoni i-tweet?
   2. Yenza enye iprojekthi yophando oya kukhetha ngayo ukuba "Yintoni oyenzayo?" Chaza ukuba kutheni.
   3. Yenza enye iprojekthi yephando apho ungathanda khona ukukhawuleza "Yintoni eyenzekayo?" Chaza ukuba kutheni.

5. [U ] "I-retweets" isoloko isetyenziselwa ukulinganisa impembelelo kunye nokusasazeka kwefuthe kwi-Twitter. Ekuqaleni, abasebenzisi bekufuneka bakopishe baze banamathele i-tweet abayithandayo, batha umlobi oyimvelaphi ngokuphatha kwakhe, kwaye uthayiphe ngesandla "RT" ngaphambi kokuba i-tweet ibonise ukuba yi-retweet. Emva koko, ngo-2009, i-Twitter yongeze iqhosha elithi "retweet". Ngomhla ka-Juni 2016, i-Twitter yenza ukuba abasebenzisi bakwazi ukubuyisa ii-tweets zabo (https://twitter.com/twitter/status/742749353689780224). Ngaba ucinga ukuba olu tshintsho lunokuchaphazela indlela oyisebenzisa ngayo "i-repubu"? Kutheni okanye kutheni?

6. [U , U , U , U ] Kwiphepha elixutyushwa ngokubanzi, uMichel kunye noogxa (2011) bahlalutye umxholo weencwadi ezingaphezu kwezigidi ezi-5 ezinokukhangela iinzame zenkcubeko. Idatha awayeyisebenzisileyo ngoku ikhutshwe njenge-Google NGrams yedatha, kwaye ngoko sinokusebenzisa idatha ukuze iphinde iphinde iphandise enye yomsebenzi wayo.

   Kwisinye seziphumo ezininzi kwiphepha, uMichel kunye nabo basebenzisana nabo bathi sikhumbula ngokukhawuleza nangokukhawuleza. Ngomnyaka othile, bathi "1883," babala inani le-1-grams epapashwe ngonyaka ngamnye phakathi kwe-1875 no-1975 "yayingu-1883". Bacinga ukuba le nxalenye ngumlinganiselo wenzala kwimicimbi eyenzeka ngaloo nyaka. Kulo mfanekiso we-3a, baceba ukulandelwa kweendlela zokusebenzisa iminyaka emithathu: 1883, 1910, no-1950. Le minyaka emithathu ibelana nomzekelo oqhelekileyo: ukusetyenziswa kancinci ngaphambi kwaloo nyaka, emva koko, ukubola. Emva koko, ukulinganisa izinga lokubola kwonyaka ngamnye, uMichel kunye nabo basebenzisana babalwa "isiqingatha-bomi" bonyaka ngamnye iminyaka yonke phakathi ko-1875 no-1975. Kulo mfanekiso 3a (ngaphakathi), babonisa ukuba isiqingatha somntu ngamnye unyaka uyancipha, kwaye baphikisana ukuba oku kuthetha ukuba silibala ngokukhawuleza ngokukhawuleza nangokukhawuleza. Basebenzisa i-Version 1 ye-English corpus, kodwa emva koko i-Google ikhuphe iinguqu yesibili ye-corpus. Nceda ufunde zonke iinxalenye zombuzo ngaphambi kokuba uqale ukubhala ikhowudi.

   Lo msebenzi uza kukunika uvavanyo lokubhala ikhowudi ehlaziywayo, iziphumo zokutolika, kunye nokudibanisa kwedatha (njengokusebenza ngeefayile ezingahambisani nokuphatha idatha ekhoyo). Lo msebenzi uza kukunceda uvuke kwaye usebenze kunye nedatha yeetata.

   1. Fumana idatha eluhlaza kwiwebhusayithi yeGoogle Books NGram Viewer. Ngokukodwa, kufuneka usebenzise inguqulelo yesi-2 yesiNgesi corpus, eyakhishwa ngoJulayi 1, 2012. Ingaphepheli, le fayili ngu-1.4GB.

   2. Phinda ufumane inxalenye ephambili yomfanekiso 3a Michel et al. (2011) . Ukuphinda ubhale kwakhona lo mfanekiso, uya kufuna ezimbini iifayile: lowo ulandelwe kwinxalenye (a) kunye nefayile "yokubala", ongayisebenzisa ukuguqula ubuninzi bokubala kwiiyure. Qaphela ukuba iifayile zeefayili zineziko ezingenza kube nzima ukuzifunda. Ingaba i-2 yesi-data ye-NGram ivelisa iziphumo ezifanayo nalabo abanikwe Michel et al. (2011) , ezisekelwe kwidatha ye-1?

   3. Ngoku khangela igrafu yakho malunga nesigrafu esakhiwe yi-NGram Viewer.

   4. Phinda ubhale umfanekiso we-3a (umzobo ophambili), kodwa utshintshe i- \(y\) -axis ukuba yi-count count mention (kungekhona izinga lokuthetha).

   5. Ngaba umehluko phakathi (b) kunye (d) unokukhokelela ukuba ubuyekeze kwakhona nayiphina imiphumo kaMichel et al. (2011). Kutheni okanye kutheni?

   6. Ngoku, usebenzisa i-number of statements, phinda ulandelelwano lwe-3a. Okokuthi, unyaka ngamnye phakathi kuka-1875 no-1975, ukubala isiqingatha sobomi baloo nyaka. Isiqingatha-ubomi sichazwa ukuba yinani leminyaka elidlulayo ngaphambi kokuba inani lokuthetha lifikele kwisiqingatha sayo sexabiso. Qaphela ukuba Michel et al. (2011) yenza into eyinkimbinkimbi ukuqikelela isigxina-sibone kwicandelo III.6 leNkcukacha yeNgcaciso ye-Intanethi-kodwa bathi iimbambo zombini zivelisa iziphumo ezifanayo. Ingaba i-2 yesi-data ye-NGram ivelisa iziphumo ezifanayo nakwezo zichazwe Michel et al. (2011) , ezisekelwe kwidatha ye-1? (Inqaku: Musa ukumangaliswa ukuba ayifuni.)

   7. Ingaba kukho naziphi na iiminyaka ezazisetyenziswa njengeminyaka ebeyilibele ngokukhawuleza okanye ngokukhawuleza? Khawucinge ngokufutshane ngezizathu ezinokwenzeka zelo phethini kwaye uchaze indlela oye wachonga ngayo ama-outliers.

   8. Ngoku uphinda uphendule esi siphumo kwi-version 2 yedatha yeNgrams ngesiTshayina, isiFrentshi, isiJamani, isiHebhere, isiTaliyane, isiSerbia nesiSpanish.

   9. Ukuthelekiswa kuzo zonke iilwimi, ngaba kukho nayiphi na iminyaka eyayingaphandle, njengeminyaka eyayilibala ngokukhawuleza okanye ngokukhawuleza? Cinga ngokufutshane malunga nezizathu ezinokwenzeka zelo phethini.

7. [U , U , U , U ] I- Penney (2016) ihlolisise ukuba ukupapashwa kwe-NSA / i-PRISM ngokubanzi (oko kukuthi, ukuvezwa kwe-Snowden) ngo-Juni 2013 kwadibaniswa nokukhawuleza ngokukhawuleza kwendlela eya kwi-Wikipedia amanqaku ngezihloko eziphakamisa iingxaki zobumfihlo. Ukuba kunjalo, olu tshintsho ekuziphatheni luya kuhambelana nempembelelo yokutshabalalisa ngenxa yokubekwa kweliso. Iindlela Penney (2016) ngamanye amaxesha zibizwa ngokuba yindlela yokuphazamiseka kwenkcazo yexesha , kwaye ihambelana neendlela ezichazwe kwicandelo 2.4.3.

   Ukukhetha isihloko esiyinhloko, uPenney okubhekiselelwe kuluhlu olusetyenziswe liSebe lase-US lokuKhuseleko lwaseKhaya ukulandelela nokubeka esweni iindaba zoluntu. Uluhlu lwe-DHS luhlula amagama athile okukhangela kwimibandela yemiba, oko kukuthi, "Ukukhathazeka kweMpilo," "ukhuseleko lweziseko," kunye ne "Ugandaganda." Kwinqanaba lokufunda, uPenney wasebenzisa amagama angama-48 anxulumene ne "Ulwahlulo" (jonga i-appendix table 8 ). Emva koko wabambisana nombono we-Wikipedia kwinqanaba leenyanga ezingama-48 ze-Wikipedia malunga neenyanga ezingama-32, ukususela ngoJanuwari 2012 ukuya ekupheleni kuka-Agasti 2014. Ukuqinisa ingxabano yakhe, naye wadala amaqela amaninzi ngokulandelela Iimbono zembali kwizinye izihloko.

   Ngoku, uya kuphinda uphinde ukwandise Penney (2016) . Yonke idatha eluhlaza oyifunayo kulo msebenzi ifumaneka kwi-Wikipedia. Okanye ungayifumana kwi-R-pack wikipediatrend (Meissner and R Core Team 2016) . Xa ubhala iimpendulo zakho, nceda uqaphele ukuba yeyiphi imithombo yedatha oyisebenzisileyo. (Qaphela ukuba lo msebenzi ubonakala kwisahluko 6.) Lo msebenzi uza kukunika ubungqina kwiinkcukacha zokumangalisa nokucinga malunga nokuhlolwa kwemvelo kwimithombo emikhulu yedatha. Kuya kukwazisa kunye nokusebenza kunye nomthombo wolwazi olunomdla kwiiprojekthi ezizayo.

   1. Funda Penney (2016) uze uphinde uphinde uguqule umfanekiso wakhe wesi-2 obonisa ukujonga kwephepha "UbuGebengu" kumaphepha anxulumene ngaphambi nangemva kweembonakaliso ze-Snowden. Chaza i ziphumo.
   2. Emva koko, phendule umhlathi 4A, othelekisa iqela lokufunda ("UbuGebengu" -amanqaku athile) kunye neqela lokuthelekisa usebenzisa amagama angundoqo ahlulwe phantsi kwe "DHS kunye nezinye iiAgesi" kwiluhlu lwe-DHS (jonga i-appendix table 10 neyokubhaliweyo 139). Chaza i ziphumo.
   3. Kwinqanaba (b) uqhathanisa iqela lokufunda kunye neqela elifanayo. U-Penney naye uthelekise namanye amaqela amabini afanisa: "Amanqaku anxulumene nezoKhuseleko" anxulumene neetafile 11) kunye namaphepha adumile e-Wikipedia (i-appendix table 12). Yiza kunye neqela elithelekisayo, kwaye uvavanye ukuba iziphumo ezivela kwinqanaba (b) lijongene nokukhetha kwakho iqela lokufanisa. Yiyiphi ikhethelo eyenza ingqiqo? Ngoba?
   4. U-Penney uthe amagama angundoqo aphathelene "nobutsha" asetyenziselwa ukukhetha amanqaku e-Wikipedia kuba urhulumente wase-US ukhankanya ubutshabungqina njengesisombululo esisisiseko sendlela yokujonga i-intanethi. Njengengheke yee-48 "Ugqirha" -izinto eziphambili, Penney (2016) naye wenza uphando kwi-MTurk, kubuza abaphenduli ukuba bahlawule i-ht yegama elilodwa ngokubhekiselele kwiNkathazo kaRhulumente, Ubuninzi -Ubuthathaka, kunye nokuPhepha (isithasiselo setafile 7 no-8 ). Phinda uphendule uphando kwi-MTurk uze uqhathanise neziphumo zakho.
   5. Ngokusekelwe kwiziphumo kwinqanaba (d) nokufunda kwakho kwinqaku, ngaba uyavumelana noPenney ukhetha isihloko sesihloko kwiqela lokufunda? Kutheni okanye kutheni? Ukuba akunjalo, ungathini na?

8. [U ] I- Efrati (2016) yabika, ngokusekelwe kwingcaciso eyimfihlo, ukuba "ukwabelana ngokupheleleyo" kwi-Facebook kuye kwayeka ngo-5.5% ngonyaka kunyaka "ukwabelana kwangaphambili" kwahla kwe-21% ngonyaka. Ukuhla kwehla kwakunzima kakhulu nabasebenzisi be-Facebook abangaphantsi kweminyaka engama-30 ubudala. Ingxelo ibonisa ukuhla kwezinto ezimbini. Omnye kukukhula kwinani labahlobo "abahlobo" kwi-Facebook. Enye into yokuba omnye umsebenzi wokwabelana uye watshintshela kumyalezo kunye nakwiinkampani ezikhuphisana njengeSpapchat. Ingxelo kwakhona yabonisa iincwadana ezininzi ze-Facebook ziye zazama ukukhulisa ukwabelana, kuquka i-News Feed ye-algorithm tweaks eyenza izikhundla zangaphambili zivelele, kwakunye nezikhumbuzo zesikhokelo sokuqala nge-"Kulo suku". Ziziphi iimpembelelo, ukuba zikhona, ngaba ezi zinto zifunyenwe ngabaphandi abafuna ukusebenzisa i-Facebook njengomthombo wedatha?

9. [U ] Uwuphi umehluko phakathi kwenzululwazi kunye nomlando-mlando? Ngokutsho kweGoldthorpe (1991) , umlinganiselo ophezulu ulawula ukuqokelela idatha. Izazi-mlando zinyanzeliswa ukuba zisebenzise iimpahla, kanti i-sociologists ziyakwenza ukuqokelela idatha yazo kwiinjongo ezithile. Funda i- Goldthorpe (1991) . Ulwahlula njani phakathi kweentlalo kunye nembali ehambelana nombono we-custommades kunye nama-readymades?

10. [U ] Oku kwakha kwi-quesiton yangaphambili. Goldthorpe (1991) yabangela iimpendulo eziliqela, eziquka enye evela kuNicky Hart (1994) eyayijongene nokuzinikela kukaGoldthorpe ukulungiswa kwedatha. Ukucacisa ukukhawulelwa kwamandla enziwe ngedatha eyenziweyo, iHart ichaza iProjekthi esebenzayo, i-survey enkulu ukulinganisa ubudlelwane phakathi kweklasi yentlalo kunye nokuvota okwenziwa yiGoldthorpe kunye noogxa phakathi ko-1960. Njengoko umntu angalindela kwi-scholar eyithandayo idatha edalwe ngedatha, i-Project Worker Project yaqokelela idatha efanelwe ukujongana neengcamango ezikhankanywe kutshanje malunga nekamva lezenhlalakahle ngexesha lokunyusa izinga lokuphila. Kodwa, i-Goldthorpe kunye nabalingane nabo "balibale" ukuqokelela ulwazi malunga nokuziphatha kokuvota kwabasetyhini. Nantsi indlela uNicky Hart (1994) atshwankathela ngayo isiqhelo sonke:

    "... kunzima ukuthintela isigqibo sokuba abesetyhini bashiywe ngenxa yokuba le nto iyenziwe" i-dataset "yayigcinwe ngumongo wepaldigmatic engabandakanyi namava amabhinqa. Eqhutywe ngumbono wemfundiso yeklasi kunye nolwaphulo njengendoda enomdla ..., i-Goldthorpe kunye nabalingane bakhe bakha isicatshulwa sobungqina bokuba bondla kwaye bazinyamekela iingcinga zabo kunokuba baveze ukuba bavavanywe ngokufanelekileyo. "

    Hart yaqhubeka:

    Iziphumo ezifunyenweyo zeProjekti YabaSebenzi abaThutheleyo zisitshelisa ngokumalunga nemilinganiselo ye-masculinist ye-sociology yenkulungwane kunokuba ikwazisa iinkqubo zokucwangcisa, ezopolitiko kunye nezinto eziphathekayo. "

    Ngaba unokucinga ngeminye imizekelo apho iqoqo leenkcukacha ezenziwe ngolu hlobo luneengxaki zokuqokelela idatha? Oku kuqhathaniswa njani ne-algorithmic confounding? Ziziphi iziphumo ezinokuba nazo xa abaphandi kufuneka basebenzise i-readymades kwaye xa kufuneka basebenzise i-custommades?

11. [U ] Kule sahluko, ndichasene nedatha eqokelelwa ngabaphandi kubaphandi abaneengxelo zolawulo ezenziwe ngamashishini noorhulumente. Abanye abantu bathi le ngxelo yezolawulo "ifunyenwe idatha," eyahlukileyo kunye ne "idatha edalwe." Kuyinyaniso ukuba iirekhodi zolawulo zifunyaniswa ngabaphandi, kodwa zenzelwe kakhulu. Ngokomzekelo, iinkampani zobugcisa zanamhlanje zisebenza nzima kakhulu ukuqokelela kunye nokucwangcisa idatha yazo. Ngaloo ndlela, ezi rekhodi zolawulo zombini zifunyenwe kwaye zenzelwe, zixhomekeke kwimiba yakho (umfanekiso 2.12).

    

    Umzobo 2.12: Umfanekiso ubini idada nomvundla; oko ubona kuxhomekeke kwindlela ojonga ngayo. Imithombo emininzi yolwazi ifumaneka kwaye yenzelwe; kwakhona, oko ubonayo kuxhomekeke kwimbono yakho. Ngokomzekelo, iirekhodi zedatha yeefowuni eqokelelwe yinkampani yeselula ifumaneka idatha kwindlela yomphandi. Kodwa, ezi ngxelo ezifanayo zenzelwe idatha ukusuka kwimbono yomntu osebenza kwisebe lenkokhelo yenkampani yefowuni. Umthombo: I-Science Monthly Popular (1899) / Wikimedia Commons .

    Ukubonelela ngomzekelo wemithombo yedata apho kubonwa kokubili kufumaneka kwaye kulungiselelwe kuyakunceda xa usebenzisa loo mthombo wedatha wophando.

12. [U ] Kwisicatshulwa esicatshulwayo, umKristu waseSandvig no-Eszter Hargittai (2015) bahlula uphando lwedijithali kumacandelo amabini abanzi ngokuxhomekeke ekubeni i-digital system "isisitye" okanye "into yokufunda." Umzekelo wohlobo lokuqala-apho inkqubo isicatshulwa nguBangtsson kunye noogxa (2011) ekusebenziseni idatha yefowuni yeselula ukulandelela ukufuduka emva kwenyikima eHaiti ngo-2010. Umzekelo woluhlobo lwesibini-apho inkqubo yinto yophando-uphando nguJensen (2007) malunga nendlela ukufakwa kweefowuni zomnxeba kulo lonke elaseKerala, e-Indiya kwathintela ukusebenza kweentengiso kwiintlanzi. Ndiyifumanisa ukuba lo mdahluko uncedo kuba ucacisa ukuba izifundo zokusebenzisa imithombo yedatha yedijithali zinokuba neenjongo ezahlukeneyo nokuba zisebenzisa uhlobo olufanayo lomthombo wedatha. Ukuze uqhubeke ucacise le ngcaciso, chaza izifundo ezine ozibonayo: ezimbini ezisebenzisa inkqubo yedijithali njengesixhobo kunye nabini abasebenzisa inkqubo yedijithali njengento yokufunda. Ungasebenzisa imizekelo kule sahluko ukuba ufuna.