InoteveraGaidhi rinotevera
Kudzokorodza neBootstrap
Tekinoroji
Nhungamiro yehunyanzvi
Kirasi Kusaenzana uye Resampling
Kusaenzana kwekirasi ndiko kana imwe mhedzisiro inodarika imwe - senge 99.9% kutengeserana kuri pamutemo maringe ne0.1% yekubiridzira - iyo inonyengedza mamodheru mukuregeredza kirasi isingawanzo asi yakakosha.
Teerera kugwaro iriAI-yakarondedzerwa · ~2 min
Pa peji ino2 min verenga
Pfupiso
Resampling inoenzanisa data rekudzidziswa kuitira kuti modhi inyatso dzidza kuona vashoma.
Kudzika Kwakadzika
Kana makirasi akatsvetwa, modhi inogona kurova 99.9% kunyatsoita nekugara uchifanotaura ruzhinji uye kusambobata hunyengeri humwechete, izvo zvisingabatsiri. Resampling inogadzirisa kugoverwa kwekudzidziswa munzira mbiri dzakafara. Kuwedzeredza makopi kana kubatanidza mienzaniso midiki - yekirasi SMOTE (Synthetic Minority Over-sampling Technique) inogadzira mapoinzi matsva nekupindirana pakati pevashoma sampuli nevavakidzani vayo vashoma pane kuvakopa. Kusampling pachinzvimbo kunorasa mienzaniso yakawanda (zvisina kurongeka, kana zvine hungwaru kuburikidza nenzira dzakaita seTomek links kana NearMiss) kuburitsa zvinhu kunze, nemubhadharo wekurasa data. Dzimwe nzira dzinodzivirira kubata data dzinosanganisira kuyera kwekirasi (kuranga zvikanganiso zvidiki zvakanyanya mukurasikirwa basa) uye kugadzirisa chikumbaridzo chesarudzo mushure mekudzidziswa.
Kuitwa muNyika yeChokwadi
Kudzidzisa dhizaini yekubiridzira kadhi rechikwereti uko hutsotsi hwechokwadi huri pasi pe1% yekutengeserana, uchishandisa SMOTE kukwidziridza nyaya dzehutsotsi husingawanzo.
Kuvaka modhi yekurapa yechirwere chisingawanzo kuwanikwa mune vashoma muzana chete yevarwere, kushandisa uremu hwekirasi saka nyaya dzakapotsa dzinorangwa zvakanyanya.
Kuona zvinhu zvakakanganisika pamutsetse wekugadzira uko zvinenge zvese zvigadzirwa zvinopfuura kuongororwa, kusampling izvo 'zvakanaka' zvinhu kuenzanisa kudzidziswa.
Kuratidzira zvisingawanzo kupindira kwetiweki mu cybersecurity logs inotongwa neyakajairwa traffic, yakaongororwa nePrecision-Recall AUC pachinzvimbo chechokwadi.
Ramangwana reKirasi Imbalance uye Resampling
Resampling iri kuwedzera otomatiki mukati meML mapaipi, aine maraibhurari senge kusaenzana-dzidza kubatanidza zvakananga mukuyambuka-kusimbisa. Tsvagiridzo iri kuenda kune inodhura-isingatarisirwe kudzidza uye akarongedzerwa kurasikirwa mabasa - senge kutarisisa kurasikirwa, uko kudzika-huremu nyore ruzhinji mienzaniso - iyo inowanzopfuura yakashata kuenzanisa pane yakadzika network. Kune tabular uye data yemifananidzo, mamodheru anogadzira anosanganisa echokwadi vashoma masamples ari kubuda semutsivi ane hunyanzvi hwekududzira SMOTE-maitiro.
Kuzvishandisa
Nei zvichikosha
- Mutengo uye bhajeti. Zvisarudzo zvezvivakwa zvinotyaira kuita uye mutengo wekushandisa kwemakore.
- Sarudzo dzakajeka. Dzidzo yehunyanzvi inobatsira zvikwata kusarudza murwi wakakodzera, kwete iwo mutsva chete.
- Kudzora kwemhando yepamusoro. Sarudzo dzeinjiniya dziri nani dzinoderedza zviitiko zvekuvimbika mukugadzira.
Chenjera
- Kugadzirisa imwe bhenji kunogona kuvanza yakafara system kushaya simba.
- Infrastructure uye mari yekugadzirisa inowanzotarisirwa pasi.
- Chengetedzo uye kucherechedzwa mapundu anogona kukura sezvo masisitimu anowedzera kuoma.
Matanho anotevera
- Tsanangura latency, mhando, uye mutengo zvinangwa usati waitwa.
- Benchmark pasi pechokwadi mutoro uye data mamiriro.
- Chishandiso chekutarisa zvikanganiso, kudonha, uye mushandisi maitiro.
- Gadzirira nzira dzekudzosera kumashure uye dzezviitiko usati wawedzera.
Ramba Uchiongorora
Tsamba yemahara
Tora iyo yemazuva ese AI pfupiso
Nhatu dzakasimbiswa dzeAI nyaya mangwanani ega ega pakati pevhiki, dzakanyorwa neChirungu chakajeka. Yemahara zvachose, hapana ads.
Imwe email zuva rega rega revhiki. Bvisa kunyoresa kamwe chete. Hatimbotengese kana kugovera kero yako.
Zviedze iwe pachako
Take the Class Imbalance and Resampling quiz
Mhinduro yepakarepo pamhinduro yega yega, uye chitupa chinogovaniswa chine ID inobatika kana wapedza kosi.
Tsigira dzidzo yemahara yeAI. AI Understanding ndeye 501(c)(3) isingabatsiri - hapana ads, hapana paywall, zvachose. Ita chipo
Mibvunzo inowanzo bvunzwa
Chii chinonzi Kirasi Imbalance uye Resampling?
Kusaenzana kwekirasi ndiko kana imwe mhedzisiro inodarika imwe - senge 99.9% kutengeserana kuri pamutemo maringe ne0.1% yekubiridzira - iyo inonyengedza mamodheru mukuregeredza kirasi isingawanzo asi yakakosha. Resampling inoenzanisa data rekudzidziswa kuitira kuti modhi inyatso dzidza kuona vashoma.
Nei kunyatsojeka kuri metric yakashata kune yakanyanya kusaenzana kurongedza dambudziko?
Kana 99.9% yezviitiko zvisina kunaka, muenzaniso unogara uchifanotaura zvisina kunaka unowana 99.9% kunyatsoita uchibata zero positives, saka chokwadi chinovanza kutadza zvachose pakirasi isingawanzo.
Chii chinoita SMOTE?
SMOTE (Synthetic Minority Over-sampling Technique) inogadzira mienzaniso mitsva mishoma nekupindirana pakati penzvimbo yevashoma nevavakidzani vayo vashoma vari pedyo, pane kungozvidzokorora.
Ndeipi nzira inobata kusaenzana PASI kushandura huwandu hwemienzaniso yekudzidziswa?
Kurema kwekirasi kunosiya iyo data isina kubatwa uye pachinzvimbo inoita kuti basa rekurasikirwa rirange kukanganisa pakirasi yevadiki zvakanyanya.
Ndeipi njodzi yakakosha yekushandisa oversampling usati wapatsanura data rako muchitima uye bvunzo seti?
Kudzokorora kusati kwapatsanurwa kunoita kuti padhuze-zvakafanana zvibodzwa zvidiki zvioneke muzvitima zvese nezviyedzo seti, zvinoburitsa ruzivo uye zvichiburitsa tarisiro yakawandisa, isingaite.
Chii chakanyanya kuderera chezvisina kurongwa undersampling?
Undersampling zviyero makirasi nekurasa mienzaniso yakawanda, iyo inogona kurasa ruzivo rweruzivo uye kukuvadza kugona kwemuenzaniso kudzidza kirasi yeruzhinji.
Ramba uchidzidza
Mazano anoenderana
Mamwe madhairekitori akasarudzirwa nyaya iyi