InoteveraGaidhi rinotevera
Kirasi Kusaenzana uye Resampling
Tekinoroji
Nhungamiro yehunyanzvi
Bootstrap resampling inofungidzira kusavimbika nekudzokorodza kudhirowa masampuli matsva, nekutsiva, kubva kune iyo data yawakaona uye kuverengazve nhamba.
Teerera kugwaro iriAI-yakarondedzerwa · ~3 min
Inogona kubatsira kutsanangura kuti chirevo kana modhi-yekuongorora metric inosiyana sei, kunze kwekunge dhizaini yekudzokorora ichienderana nemabatiro azvinoitwa.
Nhamba yakaverengerwa kubva pane imwe dataset ingori fungidziro. Imwe sampuli kubva muhuwandu humwe chete inogona kuburitsa chirevo chakasiyana, kuwirirana kana chiyero chekuongorora. Iyo bootstrap inofungidzira zvinhu zveiyo sampling mutsauko uchishandisa zvakaonekwa zviripo. Mune yakajairwa nonparametric bootstrap yekutarisa yakazvimirira, dhirowa saizi yehukuru hwepakutanga nekutsiva. Imwe neimwe dhirowa inogona kusarudza chero yepakutanga yekutarisa zvakare, saka zvimwe zvinoonekwa zvinoonekwa kanopfuura kamwe uye zvimwe zvinosiiwa. Verenga nhamba yemuenzaniso iwoyo. Dzokorora maitiro acho kakawanda kuburitsa kugovera kwebootstrap statistics. Pamufananidzo wemasvomhu, kukosha 2, 4 uye 9 kune chirevo chezvishanu. Muenzaniso unogoneka ndewe 2, 2 uye 9, iyo inorema ingangoita 4.33. Imwe yemuenzaniso haigadziri nguva yekusagadzikana. Kusiyana kwakasiyana-siyana kwakagadzirwa zvakare kwakakodzera ndiko kunoita kuti maitiro acho abatsire. A percentile interval inotora endpoints kubva percentiles yekugovera bootstrap. Dzimwe nzira, kusanganisira iyo yakasarura-yakagadziriswa uye yakakurumidza nzira inotsigirwa neSciPy, gadzirisa kuvaka kwakasiyana. Iyo yakasarudzwa nzira uye fungidziro dzayo zvine basa, kunyanya kune diki samples kana yakaoma statistics. Kuti uwane data rakapetwa, chengetedza kuwirirana. Kana uchienzanisa fungidziro yemuenzaniso pamakesi mamwe chete, edzazve imwe neimwe nyaya nemhedzisiro yayo uye zvese zvakafanotaura. Kuputsa hukama ihwohwo kunoshandura mubvunzo uri kuongororwa. Iyo bootstrap haigone kugadzira humbowo nezve vanhu kana mamiriro asipo kubva kumuenzaniso wekutanga. Izvo zvakare hazvigadzirise otomatiki kurerekera kwesarudzo kana kutsamira pakati pekutarisa. Zviyero zvakadzokororwa zvingada resampling mapoka; nguva yakatevedzana ingangoda nzira yakakodzera block. Sarudza resampling unit kuratidza magadzirirwo ekugadzira data usati wadudzira zviri pachena.
Kutanga kubva pamienzaniso yetsika 2, 4 uye 9, imwe bootstrap muenzaniso inogona kuva 2, 2 uye 9. Iine saizi yakafanana neyekutanga sampuli, inosanganisira kudzokororwa kwekutarisa uye ine chirevo chegumi netatu chakakamurwa netatu.
Kuti uenzanise mamodheru maviri pamakesi ebvunzo mamwe chete, muongorori anoedzazve zviziviso zvemakesi uye anotakura fungidziro dzemhando mbiri uye mhedzisiro yechokwadi pamwechete. Imwe neimwe resample inobereka mutsauko muzvikoro zvavo.
Chidzidzo chine marekodhi akati wandei pamunhu ari kutora chikamu. Muongorori anofunga kudzokorora vatori vechikamu vese pane kubata rekodhi yega yega seyakazvimirira.
Chikwata chinoshandisa scipy.stats.bootstrap kufungidzira kusava nechokwadi muhuwandu. Inorekodha nzira yenguva uye mbeu isina kurongeka uye inotarisa kana nguva inobuda yakagadzikana zvakakwana kuti ishandiswe.
Mishumo yekuongorora inogona kuve yakakosha nekurekodha iyo sampling unit, nzira yenguva uye chiyero chekusavimbika padivi pemiganho inobuda. Muverengi anofanira kukwanisa kutaura kana nguva ichitsanangura modhi yakatarwa pamakesi matsva kana maitiro anosanganisira kugadzirisa zvakare. Matimu anogona zvakare kuchengetedza akadzokororwa indices uye software shanduro kana reproducibility nyaya. Iyo inotevera kunatsiridza kazhinji iri nani yekuongorora sampuli pane yakawanda bootstrap kudzokorora. Kunzwisisa kuti ndeapi manyuko ekusava nechokwadi anosanganisirwa anoita kuti nguva yacho iite chiito pane kupa miganhu nhete pasina fungidziro dzavo.
Tsamba yemahara
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
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
Bootstrap resampling inofungidzira kusavimbika nekudzokorodza kudhirowa masampuli matsva, nekutsiva, kubva kune iyo data yawakaona uye kuverenga zvakare nhamba. Inogona kubatsira kutsanangura kuti chirevo kana modhi-yekuongorora metric inosiyana sei, kunze kwekunge dhizaini yekudzokorora ichienderana nemabatiro azvinoitwa.
Sample bootstrap ine saizi yepakutanga uye inodhonzwa nekutsiviwa, ichibvumira kudzokororwa kutarisa.
Idzo nhatu dzakadzokororwa tsika dzinosvika gumi nenhatu, saka kureva kwavo i13 yakakamurwa nevatatu, ingangoita 4.33.
Kuchengetedza kese-level pairing inobvumira iyo sampled zvibodzwa musiyano kuenzanisa kufanotaura kwezvakaonekwa zvakafanana.
Iyo percentile nzira inoshandisa maquantiles eiyo sampled statistics, ine 2.5% yasara mumuswe wega wega panguva iyi.
Kudzokororwa kwakawanda kunoita kuti simulation igadzikane asi haipe mitsva yakazvimiririra yekutarisa kana kugadzirisa matambudziko esampling.
Ramba uchidzidza
Mamwe madhairekitori akasarudzirwa nyaya iyi
InoteveraGaidhi rinotevera
Kirasi Kusaenzana uye Resampling
Tekinoroji