Ukuncishiswa kobukhulu
Ukwehliswa kobukhulu kunciphisa idatha ukusuka kumakholomu amaningi (izici) kwehle kuye kwezimbalwa kuyilapho kugcinwa ukwakheka okubalulekile.
Uhlolojikelele
It fights the 'curse of dimensionality,' speeds up models, and lets you actually visualize complex data in 2D or 3D.
I-Deep Dive
Amasethi edatha wangempela avame ukuba namakhulu noma izinkulungwane zezici: iphikseli ngalinye esithombeni, igama ngalinye kusilulumagama, yonke inzwa emshinini. Ezikhaleni ezinobukhulu obuphezulu kangako, amaphoyinti edatha aba yingcosana futhi aqhelelene kakhulu, izilinganiso zebanga azithembeki, futhi amamodeli avame ukugcwala kakhulu umsindo. Lesi isiqalekiso sobukhulu. Ukwehliswa kobukhulu kubeka idatha kumephu yezilinganiso ezimbalwa kakhulu kuyilapho kugcinwa ubudlelwano obunenjongo. I-PCA yenza lokhu ngokulandelana ngokuthola izikhombisi-ndlela zokuhlukahluka okukhulu. I-t-SNE kanye ne-UMAP azinamugqa futhi zihamba phambili ekuvezeni amaqoqo ukuze ubonwe ngeso lengqondo. Ukunciphisa ubukhulu kususa izici ezingafuneki noma ezinomsindo, kunqamule inkumbulo nokubala, futhi kuvame ukuthuthukisa ukunemba kwemodeli yaphansi komfula ngoba kunesiginali engabalulekile yokuyidida.
I-Technical Insight
I-PCA isebenza ngokuhlanganisa i-covariance yezici nokuthola ama-eigenvector, 'izingxenye eziyinhloko,' ezikhomba ngokuhambisana nezikhombisi-ndlela zokuhlukahluka okukhulu. Ugcina izingxenye ezimbalwa eziphezulu nedatha yephrojekthi kuzo, ulahla izikhombisi-ndlela ezihlukile ezinomsindo kakhulu. I-t-SNE ne-UMAP esikhundleni salokho ziyimodeli yobudlelwano bomakhelwane: zizama ukugcina amaphuzu abesondelene ngobukhulu obuphezulu esondele kumephu ye-low-dimensional. I-UMAP yakha igrafu yamaphoyinti aseduze, okuyenza isheshe kune-t-SNE futhi ibe ngcono ekugcineni ukwakheka okubanzi komhlaba.
I-Strategic Impact
Izinqumo ezicacile
Kukusiza ukuthi uhlukanise izimangalo ezicacile zobuchwepheshe kusukela olimini lokumaketha.
Izindleko kanye nesabelomali
Ungabuza imibuzo yokusebenzisa kangcono ngaphambi kokusebenzisa imali noma isikhathi.
Ithimba kanye nokusebenza komsebenzi
Amaqembu anokuqonda okwabiwe enza izinqumo ezingcono zomkhiqizo, inqubomgomo, nokufunda.
Ikusasa Lokuncishiswa Kobukhulu
Ukwehliswa kobukhulu manje kuyisinyathelo esivamile phakathi kwamapayipi amakhulu e-AI esikhundleni somsebenzi ozimele. I-UMAP isiphenduke inkambiso enkulu ekuhloleni okushumekiwe kusuka kumamodeli amakhulu olimi nemibono, lapho onjiniyela bakhiqiza izinkulungwane zobukhulu kumephu ye-2D ukuze bahlole ukuthi imodeli ifundile ini. Lindela ukuhlanganiswa okuqinile namadeshibhodi asebenzisanayo, ukusetshenziswa okushesha kakhulu kwe-GPU kumadathasethi wemigqa eyizigidigidi, nokusetshenziswa okukhulayo emsebenzini wokutolika, lapho abacwaningi behlisa khona ukwenza kusebenze kwangaphakathi kwemodeli ukuze baqonde futhi balungise ukuziphatha kwayo.
Ukuqaliswa Komhlaba Wangempela
Ukuhlela ukushumeka kwegama noma umusho kusuka kumodeli yolimi ku-2D nge-UMAP ukuze ubone ukuthi yimiphi imiqondo amamodeli amaqoqo ndawonye
Ukucindezela izinkulungwane zezilinganiso zomusho wofuzo ngesiguli ngasinye zibe izingxenye ezimbalwa ngaphambi kokuhlanganisa izinhlobo ezincane zezifo
Ukunciphisa izici zesithombe ngaphambi kokuziphakela kumuntu ohlukanisa isigaba ukuze ukuqeqeshwa kusheshe futhi kungabi sengozini yokugcwala ngokweqile
Ukubona ngeso lengqondo ukuziphatha kwekhasimende kumakhulukhulu amamethrikhi njengecebo lokuhlakaza le-2D ukuze ubone izingxenye zemakethe ezihlukile
Izingozi & Guardrails
Amaqembu ahlukene angasebenzisa igama elifanayo ngokuhlukile, ngakho chaza ububanzi kusenesikhathi.
Amabhentshimakhi angabukeka eqinile kuyilapho ukusebenza komhlaba wangempela kungalingani.
Ukuziba ikhwalithi yedatha nezinhlelo zokuhlaziya kuvame ukudala imiphumela entekenteke.
Ukuqalisa Umhlahlandlela
Qala ngencazelo yolimi olulula yomphumela oyidingayo.
Khetha imethrikhi eyodwa yempumelelo nesimo esisodwa sokuhluleka ngaphambi kokuhlolwa.
Qalisa umshayeli omncane onedatha emele, hhayi isethi yedemo ephucuziwe.
Idokhumenti lapho Ukunciphisa Ubukhulu kusiza nalapho izindlela ezilula zingcono.
Qhubeka Uhlole
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the Dimensionality Reduction quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Umhlahlandlela olandelayo
Amawele e-Barlow kanye Nokwehliswa Kwe-Redundancy
Imibuzo evame ukubuzwa
What is Dimensionality Reduction?
Ukwehliswa kobukhulu kunciphisa idatha ukusuka kumakholomu amaningi (izici) kwehle kuye kwezimbalwa kuyilapho kugcinwa ukwakheka okubalulekile. Ilwa 'nesiqalekiso sobukhulu,' isheshisa amamodeli, futhi ikuvumela ukuthi ubone ngeso lengqondo idatha eyinkimbinkimbi ku-2D noma i-3D.
Siyini 'isiqalekiso sobukhulu'?
Ezikhaleni ezinobukhulu obuphakeme kakhulu, amaphuzu aqhelelene kakhulu futhi ukulinganiswa kwebanga kuyehla, ngakho amamodeli ayazabalaza ukuthola amaphethini futhi avame ukugcwala ngokweqile.
I-PCA inquma kanjani ukuthi iziphi izikhombisi-ndlela okufanele zigcinwe?
I-PCA ithola izingxenye eziyinhloko, izimbazo zokuhlukahluka okukhulu, futhi igcina eziphezulu ngoba ziphethe ulwazi oluningi.
Kungani i-PCA ibizwa ngokuthi 'indlela yomugqa'?
Ingxenye eyinhloko ngayinye iyinhlanganisela yomugqa yezici zoqobo, ngakho i-PCA ayikwazi ukuthwebula isakhiwo esigobile, esingaqondile ngendlela i-t-SNE noma i-UMAP engenza ngayo.
Yiziphi i-t-SNE ne-UMAP ezisebenza kahle kakhulu kuzo?
Womabili amasu angaqondile agcina amaphuzu aseduze eduze kumephu ene-dimensional ephansi, okwenza amaqoqo abonakale ku-2D noma i-3D.