InoteveraGaidhi rinotevera
Avhareji Precision uye PR-AUC
Tekinoroji
Basics GUIDE
ROC Curves uye AUC
ROC curve inoronga kuti mugadziri anopatsanura sei makirasi maviri padanho rega rega resarudzo, uye AUC inomanikidza iyo curve yese kuita nhamba imwe.
Teerera kugwaro iriAI-yakarondedzerwa · ~2 min
Pa peji ino2 min verenga
Pfupiso
Pamwe chete vanokuudza mamiriro emhando yakazvimirira pane kwaunodhirowa cutoff.
Kudzika Kwakadzika
A Receiver Operating Characteristic (ROC) curve inoronga True Positive Rate (sensitivity, on the y-axis) vachipesana neNhema Yakanaka Mwero (1 minus chaiyo, pa x-axis) paunenge uchitsvedza chikumbaridzo chechikamu kubva pa1 kusvika ku0. Chikumbaridzo chega chega chinopa poindi imwe; kudzibatanidza kunoronda curve. Modhi inoisa chese chakanaka pamusoro pese yakaipa inombundira kona yepamusoro-kuruboshwe. Iyo Area Under the Curve (AUC) inoyera iyo yakazara nzvimbo pasi pemutsara uyu, kubva pa0.5 (random guessing, the diagonal) kusvika 1.0 (yakakwana). Tsanangudzo inobatsira: AUC yakaenzana nemukana wekuti modhi yacho inowana yakasarudzika yakasarudzwa yakanaka kupfuura yakasarudzwa zvisina tsarukano. Izwi rinobva kuHondo Yenyika II radar vanoshanda vanosiyanisa chiratidzo kubva kune ruzha.
Kuitwa muNyika yeChokwadi
Kuenzanisa maviri ehutsotsi-yekuona mamodhi ebhangi neAUC yavo kuti vatore iyo inonyanya kuenzanisa kutengeserana kwehutsotsi pamusoro peiyo iri pamutemo.
Kuongorora kuongororwa kwechirwere chechirwere (semuenzaniso, kenza yekuongorora classifier) uko radiologist inoda kutengeserana nekubata dzimwe nyaya dzichipokana nearamu enhema.
Kugadzirisa chikumbaridzo chesefa uchishandisa ROC curve kuchengetedza manyepo (tsamba dzepamutemo dzakashongedzerwa se spam) dzakaderera zvakanyanya.
Kuenzanisa kiredhiti-default scoring modhi apo AUC inopfupikisa kuti inoparadzanisa sei vanokwereta vanodzorera kubva kune avo vanotadza.
Ramangwana reROC Curves uye AUC
ROC-AUC inoramba iine metric yekubika, asi varapi vanowedzera kuibatanidza nePrecision-Recall macurves e data isina kuenzana zvakanyanya, uko ROC inogona kutaridzika inonyengera tariro. Tarisira kugamuchirwa kwakakura kwechikamu cheAUC (ichitarisa kune yakaderera-yenhema-yakanaka dunhu iro rine basa nekushanda), inoteedzera mutengo uye sarudzo-curve ongororo, uye ne-sub-boka reAUC rinoshuma kune pamusoro pemapundu. Sezvo mamodheru achidya sarudzo chaidzo, calibration metrics uye AUC ichawedzera kutaurwa parutivi pane AUC chete.
Kuzvishandisa
Nei zvichikosha
- Sarudzo dzakajeka. Inokubatsira kuparadzanisa zvakajeka zvichemo zvehunyanzvi kubva mumutauro wekushambadzira.
- Mutengo uye bhajeti. Iwe unogona kubvunza zvirinani kuita mibvunzo usati washandisa mari kana nguva.
- Team uye workflow. Zvikwata zvine nzwisiso yakagovaniswa inoita zvirinani chigadzirwa, mutemo, uye sarudzo dzekudzidza.
Chenjera
- Zvikwata zvakasiyana zvinogona kushandisa izwi rimwechete zvakasiyana, saka tsanangura nzvimbo nekukurumidza.
- Benchmarks inogona kutaridzika yakasimba nepo chaiyo-yenyika kuita isina kuenzana.
- Kuregeredza mhando yedata uye zvirongwa zvekuongorora zvinowanzogadzira mhedzisiro isina kusimba.
Matanho anotevera
- Tanga netsanangudzo yemutauro wakajeka yemhedzisiro yaunoda.
- Sarudza metric imwe yekubudirira uye imwe yekutadza mamiriro usati waedzwa.
- Mhanya mutyairi mudiki ane data remumiriri, kwete demo rakakwenenzverwa.
- Nyora uko ROC Curves uye AUC inobatsira uye uko nzira dzakareruka dziri nani.
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 ROC Curves and AUC 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 ROC Curves uye AUC?
ROC curve inoronga kuti mugadziri anopatsanura sei makirasi maviri padanho rega rega resarudzo, uye AUC inomanikidza iyo curve yese kuita nhamba imwe. Pamwe chete vanokuudza mamiriro emhando yakazvimirira pane kwaunodhirowa cutoff.
Ndeapi huwandu huviri hunoitwa neROC curve kurongana?
ROC curve inoronga Yechokwadi Positive Rate (sensitivity) pane y-axis ichipokana neNhema Positive Rate (1 - hunhu) pane x-axis mhiri kwezvikumbaridzo.
Ndeipi AUC kukosha inoenderana neye classifier isingaite zvirinani pane kungofungidzira?
A random classifier anoteedza diagonal mutsara, achipa AUC ye0.5. Iyo AUC ye1.0 yakakwana, uye 0.0 zvinoreva kuti inowana chese chese kumashure chaiko.
Sei AUC ichitsanangurwa se 'chikumbaridzo-chakazvimiririra'?
AUC inopfupikisa iyo yose ROC curve, iyo inogadzirwa nekutsvaira kwese chikumbaridzo, saka haibvi pane chero yakasarudzwa yakasarudzwa cutoff.
Kune yakanyanya kusaenzana dataset, ndeipi curve inowanzo kurudzirwa padivi kana panzvimbo yeROC nekuti ROC inogona kutarisika yakanyanyisa tariro?
Pamatambudziko asingawanzo-akanaka huwandu hukuru hwechokwadi husina kunaka hunochengeta iyo False Positive Rate yakaderera, ichiwedzera ROC. Precision-Yeuka ma curves anotarisa pane yakanaka kirasi uye anoratidza mashandiro akanyanya kutendeseka.
Why can a model have a high ROC-AUC but poorly calibrated probability estimates?
The guide notes that AUC depends on rank order and is insensitive to calibration; monotonic transformations can preserve rankings while altering the probability scale.
Ramba uchidzidza
Mazano anoenderana
Mamwe madhairekitori akasarudzirwa nyaya iyi