Njodzi inonyanya kuzivikanwa ndiyo njodzi inowanzoitika mukutsvakurudza kwevanhu; yakawedzera zvinoshamisa; uye ndiyo ngozi yakaoma kwazvo kunzwisisa.
Nzira yechipiri yekutsvaga kutsvaga kwemadijitori-yehutsva inguva yengozi yehuwandu , iyo inogona kukuvadza kubva mukuzarurirwa kwemashoko (National Research Council 2014) . Mashoko anonyadzisa kubva mukuzarurirwa kwemashoko ega ega angave ehupfumi (semuenzaniso, kurasikirwa nebasa), hutano (semuenzaniso, kunyadziswa), mupfungwa (semuenzaniso, kuora mwoyo), kana kupomerwa mhosva (semuenzaniso, kusungwa nokuda kwekuita zvisiri pamutemo). Zvinosuruvarisa, zera rejitori rinowedzera dambudziko rekuziva zvinoshamisa-pane kungowedzera mashoko pamusoro pemafambiro edu. Uye njodzi yehuwandu yakaratidza zvakaoma zvikuru kunzwisisa nekugadzirisa nekuenzaniswa nezvipembenene zvaive zvakakonzera mukutsvakurudza kwevanhu vekare, sekuda kwengozi yemuviri.
Imwe nzira iyo vaongorori magariro kudukupiswa informational ngozi ndiyo "anonymization" pamusoro mashoko. "Anonymization" ndiwo muitiro vachibvisa pachena pachavo identifiers akadai zita, kero, nenhamba kubva mashoko. Zvisinei, nzira iyi ndoda kubudirira pane vanhu vakawanda vanoziva, uye iwo, chaizvoizvo, zvikuru uye tichaona shoma. Nokuda kwechikonzero ichocho, pose Ini anorondedzera "anonymization," Ndichaenda kushandisa Taridzamutauri kukuyeuchidza kuti izvi anosika nechitarisiko kusazivana asi kwete kusazivana kwechokwadi.
Muenzaniso wakajeka wekukundikana kwe "kunyora" kunobva kuma1900 ekupedzisira muna Massachusetts (Sweeney 2002) . Gwaro reInishuwarenzi yeInishuwarenzi (GIC) yaiva sangano rehurumende rinotarisira kutenga unishuwarenzi yehutano kune vashandi vehurumende vose. Kuburikidza nebasa iri, GIC yakaunganidza mazita ehutano ehutano pamusoro pezviuru zvevashandi vehurumende. Muchiedza chekutsvaga tsvakurudzo, GIC yakasarudza kusunungura zvinyorwa izvi kuvatsvakurudzi. Zvisinei, ivo havana kugovana nedhave yavo yese; pane kudaro, ivo "vakaratidza" idzi dhiyabhorosi nekubvisa ruzivo rwakadai semazita neateri. Zvisinei, vakasiya mamwe mashoko avanofunga kuti anogona kubatsira vatsvakurudzi vakadai sehuwandu hwemashoko (zip code, zuva rokuberekwa, rudzi, uye zvepabonde) uye ruzivo rwezvehutano (shanyira deta, kuongorora, nzira) (Mufananidzo 6.4) (Ohm 2010) . Zvinosuruvarisa, iyi "kunyangadza" yakanga isina kukwana kuchengetedza data.
Kuenzanisira kukanganisa kweGIC "kusarura", Latanya Sweeney-zvino mudzidzi akapedza kudzidza paMIT-akabhadhara $ 20 kuti awane marekodhi ekuvhota kubva kuguta reCambridge, kuguta reMassachusetts gavhuna William Weld. Aya marekodhi okuvhota aisanganisira mashoko akadai sezita, kero, zip code, zuva rokuzvarwa, uye zvepabonde. Icho chikonzero chokuti dhiyabhorosi yedhafa yefodya uye faira yevhoti yakagovana-zip zip code, zuva rokuberekwa, uye zvepabonde-raireva kuti Sweeney aigona kuvabatanidza. Sweeney aiziva kuti zuva rekuzvarwa kweWeld raiva raJuly 31, 1945, uye zvinyorwa zvekuvhota zvaisanganisira vanhu vatanhatu chete muCambridge nemusi iwoyo wekuzvarwa. Uyezve, kune avo vatanhatu, vatatu bedzi vaiva varume. Uye, kubva kune varume vatatu, imwe chete yakagoverwa zip code dzeWeld. Nokudaro, dhodhi yekuvhota yakaratidza kuti chero munhu ari munharaunda yezvokurapa neWeld pamwe chete zuva rokuberekwa, chikadzi, uye zip code aiva William Weld. Muchidimbu, izvi zvikamu zvitatu zvemashoko zvakapa mimwe minwe yeminwe kwaari. Kushandisa chokwadi ichi, Sweeney akakwanisa kuwana zvinyorwa zveWederal zvekurapa, uye, kumuudza pamusoro pake, uye akamutumira kopi yemabhuku ake (Ohm 2010) .
Basa ra Sweeney rinoratidzira chimiro chekutanga chekutsvaga -kutsvaga shoko kubva kumakombiyuta ekuchengetedzwa kwemakombiyuta. Mukukurwisana uku, deta mbiri dzinotarisa, uye dzisina iyo pachayo dzinobudisa ruzivo rwakanyanya, rwakabatanidzwa, uye kuburikidza nekubatanidzwa uku, ruzivo rwakananga runoratidzwa.
Mukupindura basa ra Sweeney, uye rimwe basa rakabatana, vanotsvakurudza iye zvino vanowanzobvisa ruzivo rwakawanda-zvose zvinonzi "ruzivo rwekuzvizivisa" (PII) (Narayanan and Shmatikov 2010) -kugadzirisa "kushandiswa kwemashoko." Uyezve, vatsvakurudzi vakawanda iye zvino unoziva kuti dzimwe nhoroondo-dzakadai sedhipatimendi rezvokurapa, zvinyorwa zvemari, mhinduro dzekuongorora mibvunzo pamusoro pemitemo isiri pamutemo-zvichida yakanyanya kusununguka kusunungura kunyange pashure pe "kunyadzisa." Zvisinei, mienzaniso yandinozopa inoratidza kuti vatsvakurudzi vevanhu vanoda kuchinja mafungiro avo. Sechinhu chekutanga, kuchenjera kufunga kuti data rose inogona kuzivikanwa uye data yose inogona kuve yakangwara. Mune mamwe mazwi, panzvimbo yekufunga kuti njodzi yeruzivo inoshanda kune chiduku chiduku chemapurojekiti, tinofanira kufunga kuti inoshandiswa-kune imwe nzvimbo-kumapurogiramu ose.
Zvose zviri zviviri zvekudzokorora uku zvinoratidzwa neNetflix Prize. Sezvinorondedzerwa muchitsauko 5, Netflix yakabudisa zvikwereti zvemafirimu zana zana zvakapiwa nevanenge 500 000 nhengo, uye yakave nekufona apo vanhu vanobva munyika yose vakagadzirisa maitiro ekugadzirisa iyo Netflix kukwanisa kurumbidza mafirimu. Asati asunungura deta, Netflix yakabvisa chero ruzivo rwakananga ruzivo rwekuziva, rwakadai semazita. Vakatorawo imwe danho uye vakataura zvishoma zvinyorwa mune zvimwe zvinyorwa (semuenzaniso, kuchinja zvimwe zviyero kubva ku 4 nyeredzi kusvika ku 3 nyeredzi). Vakakurumidza kuziva, zvisinei, kuti pasinei nekuedza kwavo, data yacho yakanga isati yambozivikanwe.
Kwapera vhiki mbiri mushure mekubudiswa kwemashoko, Arvind Narayanan naVitaly Shmatikov (2008) vakaratidza kuti zvaive zvichikwanisa kudzidza nezvevamwe vanhu vanofarira mafirimu. Nenzira yekuita kwavo zvakare-kugadzirisa kusangana kwakange yakafanana na Sweeney's: kuungana pamwe nemashoko maviri ezvinyorwa, imwe ine ruzivo ruzivo rwakanyanya uye hapana ruzivo rwakananga ruzivo uye runenge rwuri ruzivo rwevanhu. Chimwe nechimwe chezvinhu izvi zvinogona kuve zvakachengeteka, asi kana zvichisanganiswa, dataset yakabatanidzwa inogona kuisa njodzi yekuziva. Munyaya yeNetflix data, hezvino kuti zvingaitika sei. Fungidzira kuti ndinosarudza kugoverana mafungiro angu pamusoro pekuita uye kunyoresa mafirimu nevashandi pamwe chete neni, asi kuti ini ndinosarudza kugoverana maonero angu nezvemafirimu ezvematongerwe enyika uye zvematongerwe enyika. Vashandi vangu vashandi vangashandisa ruzivo rwandakagovana navo kuti vawane zvinyorwa zvangu muNetflix data; ruzivo rwandinogovera runogona kunge rune minwe yakasiyana-siyana yakafanana nezuva rekuzvarwa kwaWilliam Weld, zip code, uye zvepabonde. Zvadaro, kana vakawana zvinyorwa zvangu zvakasiyana-siyana munharaunda, vanogona kudzidza maonero angu pamusoro peemafirimu ose, kusanganisira mafirimu andinosarudza kusagovana. Mukuwedzera kune rudzi urwu rwokurwiswa kunotarisana nomunhu mumwe chete, Narayanan uye Shmatikov vakaratidzawo kuti zvaikwanisika kuita nekurwisa kukuru -imwe inosanganisira vanhu vazhinji- nekubatanidza dhiabhorosi yeNetflix ne data yega uye yevhidhiyo yekucherechedza iyo vamwe vanhu vakasarudza kunyorera paIndaneti Movie Database (IMDb). Zvose zviri nyore, chero ruzivo rwakasiyana nemimwe yeminwe kune munhu chaiyo-kunyange rutivi rwavo rwekuona mafirimu-rinogona kushandiswa kuvanzwisisa.
Kunyange zvazvo nhamba yeNetflix inogona kuwanikwa zvakare mune zvinorwisana kana kuti yakawanda, inogona kuonekwa seine ngozi shoma. Mushure mezvose, kuongororwa kwemafirimu kunoratidzika sekunetseka zvikuru. Kunyange zvazvo izvi zvingava zvechokwadi kune vamwe vose, kune vamwe vevanhu 500 000 vari mudasetet, kuongororwa kwemafirimu kunogona kunge kwakanyanya. Ichokwadi, mukupindurwa kwekucherechedzazve, mukadzi akabatwa nemadzimai akabatana nekirasi-action suit neNetflix. Hezvino iyo dambudziko rakaratidzwa sei mumhosva yavo (Singel 2009) :
"[M] ovie uye nhamba ye deta ine ruzivo rwe ... yakasikwa uye yakasununguka. Dhidhiyo yemabhidhiyo inoratidza kushamwaridzana komunhu weNetflix uye / kana kurwisana nezvinhu zvakasiyana-siyana zvevanhu, kusanganisira zvepabonde, chirwere chepfungwa, kupora kubva kunwa doro, uye kushungurudzwa kubva kune vamwe vana, kubatwa zvisina kunaka, kurwisana mumhuri, upombwe, nekubatwa chibharo. "
Kudzokororazve kweNetflix Prize data inoratidza zvose kuti data yose inogona kuzivikanwa uye kuti data yose inogona kuve yakanakira. Pano, unogona kufunga kuti izvi zvinongoshanda kune data inofanirwa kuva yevanhu. Zvinotoshamisa, izvo hazvisizvo. Mukupindura chikumbiro cheMutemo weRudzi rweRuzivo, Gwaro reNew York City rakabudisa zvinyorwa zvekatekisi yese kuNew York mugore ra2013, kusanganisira pikicha uye kuregera nguva, nzvimbo, uye zvikwereti (rangarira kubva muchitsauko 2 icho Farber (2015) vakashandisa zvakafanana nedzidziso kuti vaongorore dzidziso dzinokosha mubasa rezvemari). Iyi dhesi pamusoro pekufambisa tatekisi inogona kuoneka yakashata nokuti inenge isingaratidzi kupa ruzivo pamusoro pevanhu, asi Anthony Tockar akaziva kuti iyi taxi dataset zvechokwadi yaiva nemashoko akawanda ane ruzivo pamusoro pevanhu. Kuenzanisira, akatarisa marwendo ose achitanga paHustler Club-chikwata chikuru chechikwata muNew York-pakati peusiku na6 mangwanani uye ndokuwana nzvimbo dzavo dzekudzima. Tsvakurudzo iyi yakaratidzwa-muzvinyorwa-rondedzero yekero dzevamwe vanhu vaigara vaine Hustler Club (Tockar 2014) . Zvakaoma kufungidzira kuti hurumende yeguta yakanga ine izvi mupfungwa apo yakabudisa data. Kutaura zvazviri, iyo nzira inogona kushandiswa kuwana misha yepamusha yevanhu vanoshanyira nzvimbo ipi zvayo muguta-kliniki yezvechiremba, chivako chehurumende, kana sangano rezvitendero.
Iyi miviri miviri yeNetflix Prize uye yeNew York City taxi data inoratidza kuti vanhu vane ruzivo rwakakwana vanogona kukundikana kunyatsofungidzira dambudziko rekuziva mune dhesi ravanosunungura-uye zviitiko izvi hazvisi izvo chete (Barbaro and Zeller 2006; Zimmer 2010; Narayanan, Huey, and Felten 2016) . Uyezve, mumatambudziko mazhinji akadaro, dambudziko rinodhaka rinoramba riripo paIndaneti, richiratidza kuoma kwekugadzirisa kusunungurwa kwemashoko. Pamwe chete, mienzaniso iyi-pamwe nekutsvakurudza mune kombiyuta yesayenzi pamusoro pehuga-inotungamirira kumhedziso inokosha. Vatsvakurudzi vanofanira kufungidzira kuti data rose inogona kuzivikanwa uye data yose inogona kuve yakananga.
Zvinosuruvarisa, hapana sarudzo iri nyore kune chokwadi icho data rose inogona kuzivikanwa uye kuti data rose inogona kunge yakangwara. Zvisinei, imwe nzira yekuderedza dambudziko rehuwandu hwemashoko apo iwe uri kushanda nedhidhiyo ndechokusika uye kutevera chirongwa chekuchengetedza data . Ichi chirongwa chichaderedza mukana wekuti deta yako ichanonoka uye ichaderedza kukuvadza kana kuora kunogona kuitika nenzira imwe. Zvinyorwa zvezvigadziro zvekuchengetedzwa kwemashoko, zvakadai semhando yekushandiswa kwekodzero kushandiswa, ichashandurwa kwekufamba kwenguva, asi UK Data Services inobatsira kuronga zvinhu zvepurogiramu yekuchengetedza deta mumapoka mashanu ayo anodana mavhiri mashanu : mishonga yakachengeteka, vanhu vakachengeteka , kuchengetedzwa kwakachengeteka, data yakachengeteka, uye zvakachengeteka (tafura 6.2) (Desai, Ritchie, and Welpton 2016) . Hapana mumwe wevashanu safesiti mumwe nomumwe anopa zvakakwana kudzivirirwa. Asi pamwe chete vanoumba zvinhu zvine simba zvinogona kuderedza dambudziko rekuziva.
Yakachengeteka | Action |
---|---|
Mishonga yakachengeteka | Zvinogadzirisa mapurojekiti ane data kune avo vane tsika |
Vanhu vakachengeteka | Kutsvaga kunongowanikwa kune vanhu vanogona kutendwa nedhipatimendi (semuenzaniso, vanhu vakawana maitiro edzidziso) |
Data yakachengeteka | Dhiyabhorosi akaiswa pachena uye akabatanidzwa kusvika pakukwanisika |
Zvirongwa zvakachengeteka | Dhiyabhorosi akachengetwa mumakombiyuta ane zvakakodzera mumuviri (semuenzaniso, mukamuri wakavharidzirwa) uye software (semuenzaniso, kuchengetedzwa kwephasiwedi, yakavharwa) kuchengetedzwa |
Safe output | Tsvakurudzo inobvunzurudzwa inoongororwa kuitira kudzivirira pangozi kukanganisa kwevamwe |
Mukuwedzera pakudzivirira deta yako paunenge uchishandisa, danho rimwe mudanho rekutsvakurudza apo njodzi yehuwandu inonyanya kukosha kugovana kwedhete nevamwe vaongorori. Dhaka kugoverana pakati pevasayendisiti chinhu chakakosha chekuita kwesayenzi, uye inoita kuti zvikuru kufambira mberi kwezivo. Hezvino ndiyo iyo UK House of Commons yakatsanangura kukosha kwekugovana (Molloy 2011) :
"Kuwana deta zvakakosha kana vatsvakurudzi vari kubudisa, kuongorora nekuvaka pane zviitiko zvinoratidzwa mumabhuku acho. Iko kufungidzira kunofanirwa kuva iyo, kunze kwekuti pane chikonzero chakasimba zvakasiyana, data inofanira kuziviswa zvakakwana uye inoitwa pachena pachena. "
Asi, nekugovana dheta yako nemumwe muongorori, iwe ungave uri kuwedzera dambudziko rekuziva kune vatori vechikamu. Nokudaro, zvingaita sekuti kududza dhaka kunokonzera kukanganisa kukuru pakati pezvavanosungirwa kugoverana nedzimwewo masayendisiti uye musungo wokuderedza dambudziko remashoko kune vatori vechikamu. Nenzira yakanaka, dambudziko iri harisi rakaoma sezvarinoratidzika. Pane kudaro, zviri nani kufunga pamusoro pekugoverana kwedhakwa sekuwa pamwe chete nekuenderera mberi, nechinangwa chimwe nechimwe pamusoro pekuenderera mberi kwekupa misikanano yakasiyana yezvipo kumhuri uye panjodzi kune vatori vechikamu (mufananidzo 6.6).
Pane imwe yakanyanyisa, iwe unogona kugovera deta yako pasina munhu, iyo inopedza ngozi kune vatori vechikamu asi inowedzerawo kuderedza zvigaro kuvanhu. Kune rimwewo zvakanyanya, iwe unogona kusunungura nekukanganwa , uko dheti "inozivikanwa" uye yakanyorerwa munhu wese. Kuratidzira kusasunungura data, kusunungura nekukanganwa zvinopa zvipo zvepamusoro kune zvekunze uye pangozi yevatori vechikamu. Pakati pezvinhu izvi zviviri zvakanyanya kukanganisa ndezvorudzi ruzhinji rwehuwandu, kusanganisira izvo zvandichadaidza munda wakakomberedzwa . Pasi pemugwagwa uyu, deta iri kugoverwa nevanhu vanowana zvimwe zviyero uye vanobvumirana kusungwa nemimwe mitemo (semuenzaniso, kutarisa kubva kune IRB neurongwa hwekuchengetedza data). Munda wakakomberedzwa unovhura unopa zvakawanda zvekununurwa nekukanganwa pasina pangozi. Zvechokwadi, nzira yakadaro inoita mibvunzo yakawanda-ndiani anofanirwa kuva nekwaniso, pasi pemamiriro api, uye kwenguva yakareba, ndiani anofanira kubhadhara kuchengetedza uye kupurisa munda wakakomberedzwa nemasvingo, nezvimwe-asi izvi hazvikwanisiki. Kutaura zvazviri, kune huri kutoshanda minda yemapindu munzvimbo iyo vatsvakurudzi vanogona kushandisa ikozvino, seya data archive ye Inter-yunivhesiti Consortium yePolitiki neZvesuzhinji paYunivhesiti yeMichigan.
Saka, ndeipi iyo inofanira kubva pane zvaunodzidza ichinyanya kupfurikidza kusina kugovana, munda wakakomberedzwa, uye kusunungurwa uye kukanganwa? Izvi zvinoenderana nemashoko e data yako: vatsvakurudzi vanofanira kuenzanisa Kuremekedzwa kwevanhu, Kubatsira, Ruramisiro, uye Kuremekedza Mutemo neHurumende. Kuonekwa kubva pane ino maonero, kugoverana kwemashoko hakusi yakasiyana nemitemo ye conundrum; inongova imwe chete yezvinhu zvakawanda zvekutsvakurudza umo vanotsvakurudza vanofanira kuwana zvakakodzera maitiro akanaka.
Vamwe vanotsoropodza kazhinji vanopesana nedhaka yekugoverana nokuti, mumaonero angu, ivo vanoisa pfungwa pazvipingamupinyi zvayo-izvo pasina mubvunzo-chaizvoizvo uye vanozvidza kubatsirwa kwayo. Saka, kuitira kuti kukurudzire kuisa pfungwa kune zvose zvengozi uye zvibatsiro, ndinoda kupa mhizha. Gore roga roga, motokari inotungamirira zviuru zvemurufu, asi hatiedzi kuvhara motokari. Kutaura zvazviri, kudana kunodhanha kushandiswa kunenge kusina musoro nokuti kutyaira kunobatsira zvinhu zvakawanda zvinoshamisa. Pane kudaro, vanhu vanoisa mitemo kune avo vanogona kutyaira (semuenzaniso, kudiwa kwekuve imwe nguva uye kupedza dzimwe mimwe miedzo) uye kuti vanogona sei kutyaira (semuenzaniso, pasi pekumhanya kwemhepo). Sangano rinewo vanhu vanoita nekusimbisa mitemo iyi (semuenzaniso, mapurisa), uye tinoranga vanhu vanobatwa vachivaputsa. Izvozvi rudzi rwekufunga kwakanyatsogadziriswa kwevanhu kunotungamira kutyaira kunogona kushandiswawo kugoverana kwepa data. Izvi zvinoreva kuti, panzvimbo yekuita nharo dzakakwana kana kupesana nedhaka, ndinofunga kuti tichaita kufambira mberi nekufunga nezvekuti tinogona kuderedza sei njodzi uye kuwedzera zvikomborero kubva kugoverana kwemashoko.
Kuti apedzisire, njodzi yeruzivo yakawedzera zvinoshamisa, uye zvakaoma zvikuru kufanotaura uye kuverenga. Nokudaro, zvakanakisisa kufunga kuti zvose deta zvinogona kuzivikanwa uye zvingave zvisinganzwisisiki. Kuti kuderedze dambudziko rehuwandu hwemashoko paunenge uchitsvakurudza, vatsvakurudzi vanogona kugadzira nekutevera chirongwa chekuchengetedza deta. Uyezve, njodzi yekuziva haibvumiri vatsvakurudzi kugoverana nedzimwe nevamwe vasayendisiti.