2.4.1 Counting izinto
Ukubala ngokweqa Simple kungaba ezithakazelisayo uma sihlanganise umbuzo obuhle idatha okuhle.
Nakuba kukhonjiswe ngolimi oluyinkimbinkimbi, ucwaningo oluningi lomphakathi luvele lubala izinto. Ngesikhathi semininingwane enkulu, abacwaningi bangakwazi ukubala kunanini ngaphambili, kodwa lokho akusho ukuthi kufanele baqale ukubala ngokungenasici. Esikhundleni salokho, abacwaningi kufanele babuze: Yiziphi izinto okufanele zibale? Lokhu kungase kubonakale njengendaba ehluke ngokuphelele, kodwa kunamaphethini ajwayelekile.
Ngokuvamile abafundi bakhuthaza ucwaningo lwabo lokubala ngokuthi: Ngizobala into engakaze ibe nayo ngaphambili. Isibonelo, umfundi angase athi abantu abaningi baye bafunda abafuduki futhi abantu abaningi baye bafunda amawele, kodwa akekho oye wafunda amawele abahambahambayo. Ekuhlangenwe nakho kwami, leli chinga, engilibiza ngokuba nesisusa ngokungabikho , akuvami ukuholela ocwaningweni oluhle. Isizathu sokungabikho sihlobo sokuthi kukhona imbopho lapho, futhi ngizosebenza kanzima ukuyigcwalisa. Kodwa hhayi yonke imbobo idinga ukugcwaliswa.
Esikhundleni sokugqugquzela ukungabi khona, ngicabanga ukuthi amasu angcono ukubuka imibuzo yocwaningo ebalulekile noma ethakazelisayo (noma ngokubili kokubili). Zombili lezi zimiso kunzima ukuchaza, kodwa enye indlela yokucabanga ngokucwaninga okubalulekile ukuthi inomthelela olinganiswayo noma uhlinzeka ngesinqumo esibalulekile ngabenzi bomgomo. Isibonelo, ukulinganisa izinga lokungasebenzi kungabalulekile ngoba kubonisa umnotho oqhuba izinqumo zenqubomgomo. Ngokuvamile, ngicabanga ukuthi abacwaningi banombono omuhle walokho okubalulekile. Ngakho-ke, kuyo yonke ingxenye yalesi sigaba, ngizohlinzeka ngezibonelo ezimbili lapho ngicabanga ukubala kuyathakazelisa. Esimweni ngasinye, abacwaningi babengabalwa ngokungafani; Kunalokho, babalwa ngezilungiselelo ezithile ezivezile ukwaziswa okubalulekile emibonweni eyengeziwe mayelana nendlela izinhlelo zenhlalo ezisebenza ngayo. Ngamanye amazwi, okuningi okwenza lezi zenzo zokubala ezithakazelisayo akuzona idatha ngokwayo, ivela kule mibono ejwayelekile.
Isibonelo esisodwa samandla alula wokubala sivela ekufundeni kukaHenry Farber sika (2015) ngokuziphatha kwabashayeli bamatekisi baseNew York City. Nakuba leli qembu lingase lingabonakali lithandekayo, kuyisayithi lokucwaninga lokucwaninga lokuhlola imibono emibili yokuncintisana emnothweni wezomnotho. Ngenhloso yokucwaninga kukaFarber, kunezici ezimbili ezibalulekile mayelana nomsebenzi wokusebenza kwabashayeli bamatekisi: (1) umholo wabo wehora ushintshashintsha nsuku zonke, ngokusekelwe kwezinye izinto ezifana nesimo sezulu, kanye (2) nenani lamahora umsebenzi ungashintsha usuku ngalunye ngokusekelwe ezinqumweni zabo. Lezi zici ziholela kumbuzo othakazelisayo mayelana nobuhlobo phakathi kwamahora namahora asebenzayo amahora. Izithombe ze-Neoclassical ezomnotho zibikezela ukuthi abashayeli bamatekisi bazosebenza ngaphezulu ngezinsuku lapho banezinkokhelo eziphakeme zehora. Ngaphandle kwalokho, amamodeli avela ekuziphatheni kwezezimali abikezela ngokuphambene nalokho. Uma abashayeli bebeka inhloso ethile yokuthola imali-bathi $ 100 ngosuku-futhi basebenze kuze kube yilapho lokho kuhlosiwe kuhlangene khona, abashayeli bazophela ekusebenzeni amahora ambalwa ngezinsuku abazuza ngazo ngaphezulu. Isibonelo, uma ungumuntu othola ulwazi, ungase ugcine ukusebenza amahora amane ngosuku oluhle ($ 25 ngehora) namahora amahlanu ngosuku olubi ($ 20 ngehora). Ngakho-ke, ingabe abashayeli basebenza amahora amaningi ngezinsuku ezinamaholo aphezulu amahora (njengoba kunqunywe amamodeli we-neoclassical) noma amahora amaningi ngezinsuku ezineholo eliphansi lehora ngalinye (njengoba kunqunywe imodeli yezomnotho yokuziphatha)?
Ukuze uphendule lo mbuzo u-Farber uthole idatha kuwo wonke uhambo lwamatekisi oluthathwa yi-New York City cabs kusukela ngo-2009 kuya ku-2013, idatha manje etholakalayo esidlangalaleni. Lezi zindaba-eziqoqwe ngamamitha kagesi ukuthi idolobha lingafuna amatekisi ukuba basebenzise-faka ulwazi mayelana nohambo ngalunye: isikhathi sokuqala, indawo yokuqala, isikhathi sokuphela, indawo yokuphela, ukuhamba, kanye nephuzu (uma iphothi ikhokhwe ngekhadi lesikweletu) . Ukusebenzisa le datha yemitha yamatekisi, uFarber wathola ukuthi abashayeli abaningi basebenza ngaphezulu ngezinsuku lapho inkokhelo ephakeme, ehambisana nenkolelo ye-neoclassical.
Ngaphezu kwalokhu kutholakala okuyinhloko, u-Farber wakwazi ukusebenzisa ubukhulu bedatha ukuze aqonde kangcono u-heterogeneity kanye ne-dynamics. Wathola ukuthi, ngokuhamba kwesikhathi, abashayeli abasha bafunda kancane kancane ukusebenza amahora amaningi ngosuku lwezinsuku eziphakeme (isb., Bafunda ukuziphatha njengendlela ye-neoclassical yokubikezela). Futhi abashayeli abasha abaziphatha ngokufana nalabo abakuthola amathenda baningi amathuba okuyeka ukushayela amatekisi. Zombili lezi zithole ezicashile, ezisiza ukuchaza ukuziphatha okuphawuliwe kwabashayeli bamanje, kungenzeka kuphela ngenxa yobukhulu be dataset. Babengenakuthola ekutadisheni kwangaphambili ukuthi basebenzisa amaphepha amaphepha ephepha kusuka enombolweni encane yabashayeli bamatekisi esikhathini esifushane (Camerer et al. 1997) .
Ukutadisha kukaFarber kwakuseduze nesimo esihle kakhulu socwaningo olusebenzisa umthombo omkhulu wedatha ngoba idatha eqoqwe idolobha yayiseduze kakhulu nedatha uFarber ayeyoyiqoqa (umehluko owodwa ukuthi uFarber wayeyofuna idatha kuphelele inkokhelo-ihamba kanye namathiphu-kodwa idatha yedolobha ibandakanya kuphela amathiphu akhokhelwa ngekhadi lesikweletu). Nokho, idatha kuphela ayanele. Isihluthulelo socwaningo lukaFarber sasiletha umbuzo othakazelisayo kumininingwane, umbuzo onomthelela omkhulu kunalesi simiso esithile.
Isibonelo sesibili sokubala izinto sivela ekucwaningweni nguGary King, Jennifer Pan, noMolly Roberts (2013) ngokucwaninga nge-intanethi nguhulumeni waseShayina. Kodwa-ke, kulokhu, abacwaningi kwakudingeka baqoqe idatha yabo enkulu futhi kwakudingeka babhekane nokuthi idatha yabo ayiphelelanga.
INkosi kanye nabalingani bayo bashukunyiswa yiqiniso lokuthi izikhala zokuxhumana zenhlalo e-China zihlolwe yizixhobo ezinkulu zombuso okucatshangwa ukuthi zihlanganisa amashumi ezinkulungwane zabantu. Abacwaningi nezakhamizi, kodwa, abanengqondo yokuthi lezi zicensors zithatha kanjani ukuthi okuqukethwe kufanele kususwe. Izazi zeChina zinezifiso eziphikisanayo mayelana nokuthi yiziphi izinhlobo zokuthunyelwe ezingase zisuswe. Abanye bacabanga ukuthi abacwaningi bagxila ekuthunyelwe okubaluleke kakhulu kuhulumeni, kanti abanye bacabanga ukuthi bagxile ekuthunyelwe okukhuthaza ukuziphatha okuhlangene, njengemibhikisho. Ukuthola ukuthi yikuphi okulindelweyo kuqinisile kubangelwa ukuthi abacwaningi baqonde kanjani iChina kanye nabanye ohulumeni abanobuqhawe ababandakanya ukucwaninga. Ngakho-ke, iNkosi kanye nozakwabo bafuna ukuqhathanisa izithuba ezishicilelwe futhi zisuswe ngokuthunyelwe okwakushicilelwe futhi akukaze kususwe.
Ukuqoqa lezi zikhala abathintekayo amazing wobunjiniyela of kukhasa ezingaphezu kuka-1000 Chinese media social amawebhusayithi-ngamunye ezahlukene kwamakhasi-ekutholeni izikhala ezifanele, bese siphindelela lezi zikhala ukubona eyayibonakala kamuva isusiwe. Ngaphezu izinkinga ubunjiniyela evamile ezihlobene ne ngezinga elikhulu web-ukucaca, le phrojekthi wabhekana nenselele wanezela ukuthi kwakudingeka ukuba kube ngokwedlulele fast ngoba izikhala eziningi zazihlolwa zithathwa phansi emahora langaphasi kulangu-24. Ngamanye amazwi, a abakhasi slow owayezolahlekelwa inkatho izikhala zazihlolwa. Ngaphezu kwalokho, lo crawlers kwadingeka ukuba ngenze konke lokhu iqoqo lemininingwane ngesikhathi ngibalekela sokuzwa funa kumawebhusayithi ezezindaba ephathelene nabantu ukuvimbela noma ngenye beye bashintsha izimiso zabo eziphendula cwaningo.
Ngesikhathi leso msebenzi omkhulu wobunjiniyela usuqedile, iNkosi kanye nabalingani bakhe bathole izikhala ezingaba ngu-11 ngezihloko ezi-85 ezahlukahlukene ezicacisiwe, ngasinye sinomqondo wokuzwela. Isibonelo, isihloko sokuzwela okukhulu ngu-Ai Weiwei, umculi ophikisanayo; Isihloko sokuzwela okuphakathi kuwukubonga nokuqhathaniswa kwemali yaseShayina, futhi isihloko sobuzwe obuphansi yiNdebe Yomhlaba. Kulezi zikhala eziyizigidi ezingu-11, cishe izigidi ezingu-2 zacutshungulwa. Ngokumangalisa ukuthi, iNkosi kanye nozakwethu bathola ukuthi izikhala ezinhlokweni ezibucayi zithathwa ngokuphindwe kaningi kunezikhalazo ezihlokweni eziphakathi nobusuku obuphansi. Ngamanye amazwi, ama-censors aseShayina cishe angabheka okuthunyelwe okukhuluma nge-Ai Weiwei njengeposi ekhuluma ngeNdebe Yomhlaba. Lokhu okutholakele akusekeli umqondo wokuthi uhulumeni ucacisa zonke izikhala ezihlokweni ezibucayi.
Lokhu kubalwa okulula kwezinga lokucwaninga ngesihloko kungahle kudukise, noma kunjalo. Isibonelo, uhulumeni angase acacise okuthunyelwe okusekela i-Ai Weiwei, kodwa ushiye okuthunyelwe okubaluleke kakhulu kuye. Ukuze bahlukanise phakathi kokuthunyelwe ngokucophelela, abacwaningi kwakudingeka balinganise ukuzwa kweposi ngalinye. Ngeshwa, naphezu komsebenzi omningi, izindlela ezizenzekelayo zokuzithokozisa ngokusebenzisa izichazamazwi ezikhona ngaphambili namanje akuzona ezinhle kakhulu ezimweni eziningi (cabanga emuva ezinkingeni zokwakha isikhathi somzuzu ka-September 11, 2001 esichazwe esigabeni 2.3.9). Ngakho-ke, iNkosi kanye nosebenza nabo babedinga indlela yokubiza izethulo zabo zokuxhumana ezithinta izigidi ezingu-11 ukuthi ngabe (1) zibucayi yini isimo, (2) zisekela uhulumeni, noma (3) imibiko engabalulekile noma engokoqobo mayelana nezenzakalo. Lokhu kuzwakala njengomsebenzi omkhulu, kodwa bawuxazulula usebenzisa inkolelo enamandla ejwayelekile kwisayensi yezesayensi kodwa engavamile kwisayensi yenhlalo: ukufunda okuqondisiwe ; bheka umfanekiso 2.5.
Okokuqala, esiteshini esivame ukubizwa ngokuthi i- preprocessing , abacwaningi baphendulela okuthunyelwe kwezokuxhumana kwezenhlalo zibe matrix yesikhathi sedokhumenti , lapho kwakukhona umugqa owodwa wedokhumenti ngayinye nekholomu eyodwa ebhalwe ukuthi ngabe okuthunyelwe kuqukethe igama elithile (isb., Ukuphikisa noma ithrafikhi) . Okulandelayo, iqembu labasizi bezokucwaninga ngesandla-labeka umqondo wesampula kokuthunyelwe. Khona-ke, basebenzisa leli datha lwedatha ukudala umshini wokufunda umshini ongashukumisa ukuzwa kokuthunyelwe okusekelwe izici zalo. Ekugcineni, basebenzisa le model ukulinganisa ukuzwa kwazo zonke izigidi 11 izikhala.
Ngakho-ke, kunokuba ufunde ngokubhala ngesandla futhi ubhale amagama angu-11 wezigidi-okungeke kube khona okungenakwenzeka-iNkosi nezakwabo babhalwe ngesandla ngenombolo encane yokuthunyelwe bese basebenzisa ukufunda okuqondisiwe ukulinganisa ukuzwa kwazo zonke izikhalazo. Ngemva kokuqedela lokhu kuhlaziywa, bakwazi ukuphetha ngokuthi, ngokumangalisa ukuthi kungenzeka ukuthi okuthunyelwe kususwe akuhambisani nalokho okwakubaluleke kakhulu kumbuso noma ukusekela uhulumeni.
Umdwebo 2.5: Isimiso esiluliwe senqubo esetshenzisiwe King, Pan, and Roberts (2013) ukulinganisa ukuzwa kwezigidi ezingu-11 zokuthunyelwa kwezokuxhumana kweShayina. Okokuqala, e isinyathelo preprocessing, abacwaningi iguqulwe okuthunyelwe zokuxhumana ku-matrix idokhumenti eside (bheka Grimmer and Stewart (2013) ukuze uthole olunye ulwazi). Okwesibili, babhala ngesandla-isampula encane yokuthunyelwe. Okwesithathu, baqeqesha imodeli yokufunda eqondisiwe ukuze bahlukanise ukuthunyelwa kokuthunyelwe. Okwesine, basebenzisa imodeli yokufunda eqondisiwe ukulinganisa ukuzwa kwazo zonke izikhalazo. Bheka King, Pan, and Roberts (2013) , isithasiselo B ngencazelo eningiliziwe.
Ekugcineni, iNkosi kanye nabalingani bakhe bathole ukuthi izinhlobo ezintathu kuphela zokuthunyelwe zazihlaziywa njalo: izithombe zobulili ezingcolile, ukugxekwa kwabacwaningi, kanye nalabo ababenezinyathelo ezihlangene (okungukuthi, kungenzeka ukuholela emibhikisho emikhulu). Ngokubheka inani elikhulu lezethulo ezazisuswa kanye nokuthunyelwe okungasusiwe, iNkosi kanye nozakwethu bakwazi ukufunda ukuthi ama-censors asebenza kanjani ngokubuka nokubala. Ngokuqhubekayo, kufanekisela isihloko esizokwenzeka kulo lonke le ncwadi, indlela yokufunda eqondisiwe ayisebenzisayo-ukufaka ngesandla nge-labeling eminye imiphumela bese ikhiqiza imodeli yokufunda yomshini ukubiza ukuthi konke okuphumayo kuyinto evamile ekucwaningweni kwezenhlalakahle eminyakeni yobudala . Uzobona izithombe ezifana kakhulu nokulinganisa 2.5 ezahlukweni 3 (Ukubuza imibuzo) no-5 (Ukudala ukubambisana okukhulu); lena ingenye yemibono embalwa evela ezahlukweni eziningi.
Lezi zibonelo-ukuziphatha kokusebenza kwabashayeli bamatekisi eNew York kanye nokuziphatha kokucwaninga kwezokuxhumana komphakathi wase-China-kubonisa ukuthi ukubalwa okulula kwemithombo emikhulu yedatha kungaholela ekucwaningeni okuthakazelisayo nokubalulekile. Nokho, kuzo zombili izimo, abacwaningi kwakudingeka balethe imibuzo ezithakazelisayo kumthombo omkhulu wedatha; idatha ngokwayo ayanele.
