2.3.10 hankali

Wasu daga cikin bayanai da kamfanoni da gwamnatoci da shi ne m.

Kamfanonin inshora na kiwon lafiya suna da cikakkun bayanai game da lafiyar da abokan ciniki suka karɓa. Za'a iya amfani da wannan bayanin don bincike mai muhimmanci game da lafiyar, amma idan ya zama sanadiyar jama'a, zai iya haifar da mummunan rauni (misali, kunya) ko ci gaban tattalin arziki (misali, asarar aiki). Yawancin manyan manyan bayanan bayanan suna da bayanin da ke da mahimmanci , wanda shine wani dalili na dalilin da yasa basu iya yiwuwa ba.

Abin baƙin cikin shine, ya zama abin ƙyama don yanke shawara game da abin da aka sani (Ohm 2015) , kamar yadda Netflix Prize ya kwatanta. Kamar yadda zan bayyana a babi na 5, a shekara ta 2006 Netflix ya bada kyautar fim din miliyan 100 da kimanin kusan 500,000 suka samu kuma yana da kira na budewa inda mutane daga ko'ina cikin duniya suka bada algorithms wanda zai iya inganta ikon Netflix don bayar da shawarar fina-finai. Kafin sake watsar da bayanan, Netflix cire duk wani bayanan da yake bayarwa na sirri, kamar sunaye. Amma, kawai makonni biyu bayan bayanan da aka fitar da Arvind Narayanan da Vitaly Shmatikov (2008) ya nuna cewa yana yiwuwa a koyi game da fina-finan fim din mutane na musamman ta hanyar amfani da abin da zan nuna maka a babi na 6. Ko da yake wani mai tuƙi zai iya gano samfurin fim din mutum, har yanzu ba ya zama abin damuwa a nan. Duk da cewa wannan zai zama gaskiya a gaba ɗaya, domin akalla wasu mutane 500,000 a cikin dataset, sharuddan fina-finai sun damu. A hakikanin gaskiya, saboda amsawa da sake sake bayanan bayanan, mace mai lakabi ta hade ta shiga wani aiki da aka yi da Netflix. Ga yadda aka bayyana matsala a wannan shari'a (Singel 2009) :

"[M] ovie da rating data dauke da bayanin wani ... sosai sirri da kuma m yanayi. Bayanin fim na memba ya nuna sha'awar da ke da nasaba da abubuwan da ke cikin sirri, ciki har da jima'i, rashin lafiyar hankali, dawowa daga shan giya, da kuma cin zarafi, cin zarafin jiki, tashin hankalin gida, zina, da fyade. "

Wannan misali ya nuna cewa za'a iya samun bayanin da wasu ke ɗaukan hankali a cikin abin da zai iya zama alamar basira. Bugu da ari, yana nuna cewa babban tsaro da masu bincike suka yi amfani da shi don kare ƙananan bayanai-de-identification-na iya kasawa a hanyoyi masu ban mamaki. Wadannan ra'ayoyin biyu suna ci gaba a cikin cikakken bayyani a babi na 6.

Abu na ƙarshe don tunawa game da bayanai mai mahimmanci shine karɓar shi ba tare da yardawar mutane ba ya kawo tambayoyin dabi'a, koda kuwa ba'a haifar da cutar ba. Kusan kamar kallon mutum yana shan ruwa ba tare da yardarta ba za a iya la'akari da saɓin sirrin mutumin, tattara bayanai mai mahimmanci - kuma ku tuna yadda wuya zai iya yanke shawarar abin da ke damun-ba tare da izini ya haifar da damuwa ba. Zan dawo tambayoyin game da sirri a babi na 6.

A ƙarshe, manyan asusun bayanai, irin su sha'anin gwamnati da kuma harkokin kasuwanci, ba a halicce su ba ne don manufar bincike na zamantakewa. Babban mahimman bayanai na yau, da kuma gobe gobe, suna da halaye 10. Yawancin kaddarorin da ake ganin suna da kyau don bincike-manyan, da-da-da-wane, da kuma marasa aiki-sun zo ne daga gaskiyar a cikin kamfanoni na zamani da gwamnatoci suna iya tattara bayanai a wani sikelin da ba zai yiwu a baya ba. Kuma da yawa daga cikin dukiyar da aka yi la'akari da su ba daidai ba ne don bincike - bai cika ba, rashin yiwuwar, ba tare da bambanci ba, drifting, algorithmically confounded, m, datti, da kuma m-zo ne daga gaskiyar cewa masu bincike ga masu bincike basu tattara waɗannan bayanai ba. Ya zuwa yanzu, na yi magana game da gwamnati da harkar kasuwanci tare, amma akwai wasu bambance-bambance tsakanin su biyu. A cikin kwarewa, bayanai na gwamnati sun kasance marasa galibi, ba su da alamun algorithmically, da kuma rashin drifting. Ɗaya daga cikin bangaren, sha'anin aikin kula da harkokin kasuwanci sun kasance mafi yawancin lokaci. Fahimtar waɗannan halaye na gaba guda 10 shine mataki na farko da zai taimaka wajen ilmantarwa daga manyan asusun bayanai. Kuma yanzu mun juya zuwa hanyoyin bincike da za mu iya amfani da wannan bayanan.