Basics GUIDE

F1 Score uye F-Beta

F1 inosanganisa kuomarara uye kudzokorora kuburikidza nehwaro hwavo hwekupindirana, ichibatsira muenzaniso chete kana zvose zviviri zvakasimba zvakakwana.

Teerera kugwaro iriAI-yakarondedzerwa · ~3 min

Pa peji ino3 min verenga
  1. Pfupiso
  2. Kudzika Kwakadzika
  3. Kuitwa muNyika yeChokwadi
  4. Ramangwana reF1 Score uye F-Beta
  5. Kuzvishandisa
  6. Ramba Uchiongorora
  7. Mibvunzo inowanzo bvunzwa

Pfupiso

F-beta inoshandisa paramita yekuisa mutoro wakawanda pakudzokorora kana kuomarara kune basa rakatarwa. Mibairo iyi inobva pakusarudzwa kwechikamu chakanaka, muganho uye nzira yekuwedzera uye haina kusanganisa mutengo wese kana kusarura kwekugadzirisa.

Kudzika Kwakadzika

Precision asks what share of predicted positive cases are truly positive; recall asks what share of actual positive cases were found. F1 combines them as the harmonic mean, 2 × precision × recall divided by their sum when that denominator is nonzero. The harmonic mean falls when one component is low, so a classifier cannot earn a high F1 merely by excelling at one side. The score is useful for comparing systems with the same task definition, but it does not tell a team which kind of mistake matters more. Consider invented counts of eight true positives, two false positives and eight false negatives. Precision is 8/(8+2) = 0.8 and recall is 8/(8+8) = 0.5. F1 is 2 × 0.8 × 0.5/(0.8+0.5), about 0.615. The equivalent count formula is 2TP/(2TP+FP+FN), or 16/26. Report the counts as well as the score so readers understand what happened. F-beta generalizes F1. When beta is greater than one, recall carries more weight; when beta is less than one, precision does. F2 may suit a screening workflow where missing a relevant case is more costly than sending another case for review, but the choice needs an operational justification. It does not eliminate the need to inspect false positives, capacity limits or subgroup outcomes. A threshold change can trade precision against recall and therefore alter F-beta even if the underlying score model is unchanged. For multiple classes, macro averaging computes the per-class result and then gives each class equal weight. Micro averaging pools true positives, false positives and false negatives across classes before computing the score. These answer different questions when class frequencies vary. State the positive label and averaging rule explicitly. An F1 number says nothing directly about probability calibration, economic cost, or whether a small subgroup is being missed. Evaluate those separately, and use a held-out dataset that represents the decisions the model will face.

Kuitwa muNyika yeChokwadi

Chikwata chekuongorora chinoratidza F2 kana kushaikwa kwezviitiko zvakakosha kuchikosha kupfuura kuongorora chiziviso chakaipa, ichiratidza zvakare kuomarara uye kudzokorora.

Muongorori anoverenga F1 kubva pamapoinzi masere akakodzera, maviri akashata uye masere akashata ekushaya, muenzaniso wakagadzirwa.

Muongorori wemulticlass anoratidza macro F1 kuratidza maitiro ekushaya zvishoma zvakasiyana panzvimbo pekuratidza chete avhareji yose.

Chikwata chechigadzirwa chinoenzanisa metric pamatanho akati wandei ekusarudza uye chinoona kuti sarudzo yakasarudzwa inokodzera kugona kwekuongorora uye mutengo wezvikanganiso.

Ramangwana reF1 Score uye F-Beta

Sezvo ongororo dzeAI dzichiwedzera kupinda mumakirasi nemamiriro ekushandisa, nhamba imwechete yeF1 ichava isina kukwana pakuita sarudzo. Makteam anogona kushandisa F-beta kuti aone kusarudzwa kwakanyorwa kwekudzokorora kana kunyatsochera, asi kukosha kwebeta kunoramba kuri kudzikisira mutengo chaiwo. Mishumo iri nani inofanira kuratidza huwandu, mutemo wekupatsanura, muganho wekushanda uye kusava nechokwadi pamwe chete. Shanduko dzevanhu dzinogona kuchinja kunyatsochera kunyangwe kudzokorora kuri kuchengetedza, saka ma metrics anoda kuvandudzwa mushure mekuisa. Mabhodhi anotevera anogona kuita kuti kutora zvibodzwa kuve nyore kuongorora, asi kusarudza kuti ndiani anobata kukanganisa kwega kunoramba kuri mutemo uye sarudzo yebasa isingagone kugadzirisa nehurongwa hwekombiyuta.

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

  1. Tanga netsanangudzo yemutauro wakajeka yemhedzisiro yaunoda.
  2. Sarudza metric imwe yekubudirira uye imwe yekutadza mamiriro usati waedzwa.
  3. Mhanya mutyairi mudiki ane data remumiriri, kwete demo rakakwenenzverwa.
  4. Tsanangura nzvimbo apo F1 Score neF-Beta zvinobatsira uye apo nzira dziri nyore 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

Tora F1 Score neF-Beta mubvunzo

Mhinduro yepakarepo pamhinduro yega yega, uye chitupa chinogovaniswa chine ID inobatika kana wapedza kosi.

Tanga mibvunzo

Tsigira dzidzo yemahara yeAI. AI Understanding ndeye 501(c)(3) isingabatsiri - hapana ads, hapana paywall, zvachose. Ita chipo

Mibvunzo inowanzo bvunzwa

Chii chinonzi F1 Score neF-Beta?

F1 inobatanidza kunyatsoita (precision) nekuyeuka (recall) kuburikidza neharmonic mean yayo, ichipa mubairo kumudziyo (classifier) chete kana zvose zviri zviviri zvakasimba. F-beta inoshandisa parameter kuti iwedzere mutoro pakuyeuka kana kunyatsoita pabasa rakatarwa. Aya ma score anovimba nekirasi yakanaka yakasarudzwa, muganho (threshold) uye nzira yekuaverenga, uye haagumiri pakunzwisisa mari dzese kana nyaya dzekuremekedza.

F1 inobatanidza sei kunyatsoita nekuyeuka?

F1 i2PR/(P+R), harmonic mean yekunyatsoita nekuyeuka kana zvatsanangurwa.

Mumuenzaniso wakavakwa mu gwara une 8 TP, 2 FP uye 8 FN, F1 ingangoita zvakadini?

Fomula yehuwandu inopa F1 = 2TP/(2TP+FP+FN) = 16/26 ≈ 0.615.

Kana beta ikapfuura imwe muF-beta, chii chinopiwa mutoro wakawanda?

F-beta iine beta yakapfuura imwe inonyanya kusimbisa kuyeuka kupfuura kunyatsoita, pasina kubvisa kunyatsoita.

Basa rinokoshesa kudzivirira zviziviso zvenhema kupfuura kubata nyaya dzese dzakanaka. Beta ipi inogona kuratidza izvozvo?

Beta iri pakati pe zero neimwe inopa kunyatsoita mutoro wakanyanya, kunyange zvazvo mari dzechokwadi dzichiri kuda kuongororwa.

Macro F1 inoita chii pamakirasi akawanda?

Macro averaging inopa kukosha kwakafanana kune F1 yechikamu chimwe nechimwe, ichiita kuti maitiro eminority-class aoneke zvakanyanya kupfuura pfupiso yakavharirwa neprevalence.