Isku dheelitir la'aanta fasalka iyo dib u qaabaynta
Class imbalance is when one outcome vastly outnumbers another — like 99.9% legitimate transactions versus 0.1% fraud — which tricks models into ignoring the rare but important class.
Dulmar
Resampling rebalances the training data so the model actually learns to spot the minority.
quusid qoto dheer
Marka fasallada la qalloociyo, qaabku wuxuu ku dhufan karaa saxnaanta 99.9% isagoo had iyo jeer saadaaliya inta badan oo aan waligiis qaban hal khiyaano, taas oo aan faa'iido lahayn. Dib-u-samaynta waxay hagaajinaysaa qaybinta tababarka laba siyaabood oo ballaaran. Naqshadeynta xad-dhaafka ah ee nuqul ka mid ah ama isku-dubarid tusaalayaal laga tirada badan yahay - SMOTE-ga caadiga ah (Farsamaynta Muundooyinka-samaynta laga tirada badan yahay) waxay abuurtaa dhibco cusub iyadoo la dhexgelinayo muunadda beelaha laga tirada badan yahay iyo deriskeeda beelaha laga tirada badan yahay ee ugu dhow halkii ay koobi lahaayeen. Hoos-u-dhigiddu waxay meesha ka saartaa tusaalooyinka badidooda (si aan kala sooc lahayn, ama si caqli-gal ah iyada oo loo marayo hababka sida Tomek links ama NearMiss) si ay xitaa wax u dhacaan, oo ay ku kacayso kharashka tuurista xogta. Beddelka ka fogaanaya taabashada xogta waxaa ka mid ah miisaanka fasalka ( ciqaabta khaladaadka laga tirada badan yahay ee shaqada luminta) iyo hagaajinta heerka go'aanka tababarka ka dib.
Aragtida Farsamada
Xeer muhiim ah: dib u miuse kaliya shaxda tababarka, waligaa ansaxinta ama dejinta tijaabada, oo had iyo jeer dib u samee gudaha laalaabka xaqiijinta. Si xad dhaaf ah loo kala saaro ka hor inta aan la kala saarin waxa soo daadanaya dhibco laba-laaban ah oo ku jira hab-raaca imtixaanka oo buunbuunisay buundooyinka. Sababtoo ah saxnimadu halkan macno ma leh, qiimayntu waa in ay ku tiirsato saxnaanta, dib-u-celinta, F1, Precision-Recall AUC, ama Matthews Correlation Coefficient - cabbirro daacad ah marka fasalka togan uu naadir yahay.
Saamaynta Istiraatijiyadeed
Qiimaha iyo miisaaniyada
Go'aamada qaab-dhismeedku waxay horseedaan waxqabadka iyo kharashka hawlgalka sannadaha.
Go'aamo cad
Waxbarashada farsamada waxay ka caawisaa kooxaha inay doortaan xidhmo sax ah, ma aha oo kaliya kan ugu cusub.
Xakamaynta tayada
Doorashooyinka injineernimada ee wanaagsan waxay yareeyaan shilalka la isku halleyn karo ee wax soo saarka.
Mustaqbalka Isku-dheellitir la'aanta Heerka iyo Dib-u-samaynta
Dib-u-samaynta ayaa si otomaatig ah ugu sii kordheysa gudaha dhuumaha ML, oo leh maktabado sida-wax-barashada-dheellitir la'aanta ah oo si toos ah ugu biireysa ansaxinta isdhaafsiga. Cilmi-baadhistu waxay u wareegaysaa dhinaca waxbarashada xasaasiga ah iyo hawlaha khasaarinta ee ku habboon - sida luminta focal, taas oo hoos u dhigta tusaalooyin badan oo fudud - kuwaas oo inta badan ka sarreeya dib-u-qaabaynta shabakadaha qoto dheer. Xogta jaantuska iyo sawirka, moodooyinka wax-soo-saarka ee soo saara muunado laga tirada badan yahay ee dhabta ah ayaa u soo baxaya sidii kuwo aad u casrisan oo ku xiga isdhexgalka qaabka SMOTE.
Dhaqangelinta Adduunka-dhabta ah
Tababbarka baaraha khayaanada kaadhka kreditka halkaasoo khiyaanada dhabta ahi ay aad uga hoosayso 1% wax kala iibsiga, adoo isticmaalaya SMOTE si loo kordhiyo kiisaska khiyaanada ee naadirka ah
Dhisidda qaabka caafimaad ee cudur naadir ah oo ku jira kaliya dhowr boqolkiiba bukaannada, codsanaya miisaanka fasalka si kiisaska la seegay si aad ah loo ciqaabo
Ogaanshaha walxaha cilladaysan ee khadka wax-soo-saarka halkaas oo ku dhawaad dhammaan alaabtu ay dhaafaan kormeerka, iyadoo hoos loo dhigayo walxaha 'wanaagsan' si loo dheellitiro tababarka
Calaamadinta faragelinta naadirka ah ee shabakadaha diiwaannada amniga internetka oo ay u badan yihiin taraafikada caadiga ah, oo lagu qiimeeyay Precision-Recall AUC halkii ay sax ahaan lahayd
Khatarta & Dariiqyada Ilaalada
Hagaajinta hal bartilmaameed waxay qarin kartaa daciifnimada nidaamka ballaaran.
Kaabayaasha dhaqaalaha iyo dayactirka inta badan waa la dhayalsadaa.
Nabadgelyada iyo daldaloolada u fiirsashada ayaa kori kara marka nidaamyadu noqdaan kuwo aad u adag.
Qorshe Hawleedka Dhaqangelinta
Qeex daahida, tayada, iyo bartilmaameedyada qiimaha ka hor inta aan la hirgelin.
Benchmark marka la eego culeyska dhabta ah iyo xaaladaha xogta.
La socodka qalabka khaladaadka, leexashada, iyo saamaynta isticmaalaha.
U diyaari dib-u-noqoshada iyo dariiqyada jawaab-celinta dhacdada ka hor inta aanad miisaan.
Sii wad Sahaminta
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 Class Imbalance and Resampling 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
Hagaha xiga
Shabakadda Siamese iyo Khasaaraha Saddex-geesoodka ah
Su'aalaha soo noqnoqda
What is Class Imbalance and Resampling?
Isku dheelitir la'aanta fasalka waa marka hal natiijadu ay aad uga badan tahay mid kale - sida 99.9% macaamil ganacsi oo xalaal ah oo ka soo horjeeda 0.1% khiyaamo - taas oo khiyaanaysa moodooyinka in ay iska indhatiraan fasalka naadirka ah laakiin muhiimka ah. Dib-u-samaynta waxay dib u miisaamaysaa xogta tababarka si uu moodalku dhab ahaantii u barto in uu ogaado kuwa laga tirada badan yahay.
Waa maxay sababta saxnaanta cad u tahay mitir liidata ee dhibaatada soocidda aadka u dheelitiran?
Haddii 99.9% kiisasku ay taban yihiin, qaabka had iyo jeer saadaaliya taban wuxuu helayaa saxsanaanta 99.9% marka la qabanayo natiijada eber, markaa saxnaanta waxay qarinaysaa guuldarada guud ee fasalka naadir ah.
Muxuu sameeyaa SMOTE?
SMOTE (farsamo muunad ka badan oo laga tirada badan yahay) waxay abuurtaa tusaalooyin cusub oo laga tirada badan yahay iyadoo is dhexgalaysa barta beelaha laga tirada badan yahay iyo deriskeeda beelaha laga tirada badan yahay ee ugu dhow, halkii si fudud loo koobi lahaa.
Habka kee u maareeya dheelitir la'aanta IYADOO aan la beddelin tirada tusaalooyinka tababarka?
Miisaanka fasalka wuxuu ka tagayaa xogta iyada oo aan la taaban oo taa beddelkeeda waxay ka dhigtaa shaqada lumis inay ciqaabto khaladaadka fasalka laga tirada badan yahay si aad u culus.
Waa maxay khatarta muhiimka ah ee isticmaalka xad-dhaafka ah ka hor inta aanad xogtaada u kala qaybin tareenka iyo qaybaha tijaabada?
Dib-u-habaynta ka hor kala qaybsanaanta ka hor waxay u oggolaanaysaa dhibcooyin isku dhow oo isku mid ah inay ka soo muuqdaan tareenka iyo qaybaha tijaabada, daadinta macluumaadka oo soo saara rajo xad dhaaf ah, waxqabad aan macquul ahayn.
Waa maxay faa'iidada ugu weyn ee sampling-yarida random?
Faham la'aanta waxay dheellitirtaa fasallada iyadoo la tuurayo tusaalayaal badan, kuwaas oo xoori kara xogta macluumaadka waxayna dhaawici kartaa awoodda moodeelka si uu u barto fasalka aqlabiyadda.