Ikibazo gihuza ibibazo · Hagati Urwego

Ikibazo cyo Kwisubiraho

Kugenzura sigmoid na softmax bishoboka, cross-entropy ikwiranye, gusobanura log-odd ibisobanuro, imipaka ntarengwa, hamwe no guhitamo bisanzwe.

Inzira zijyanye no kuyoboraKwisubiraho
Ikibazo 1 Bya 8

Nubwo izina ryayo, gusubira inyuma kwa logistique bikoreshwa iki?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Nubwo izina ryayo, gusubira inyuma kwa logistique bikoreshwa iki?
    • Gutondekanya mu byiciro
    • Guteganya imibare ikomeza
    • Gukusanya amakuru atanditse
    • Kugabanya ibipimo
  2. Niki imikorere ya sigmoid (logistic) ikora murugero?
    • Ihitamo umubare wibiranga
    • Ikarita agaciro kose kubishoboka hagati ya 0 na 1
    • Ikuraho hanze
    • Ihuza amakuru
  3. Nibihe bikorwa byo gutakaza bigabanya gusubira inyuma mugihe cy'amahugurwa?
    • Ikigereranyo cya kare
    • Inertia
    • Gutakaza ibiti (cross-entropy)
    • Gutakaza Hinge
  4. Ni ukubera iki gusubira inyuma bihabwa agaciro kubisobanuro?
    • Ntabwo ikoresha imibare
    • Ihisha logique yimbere
    • Buri gihe iteganya 50 ku ijana
    • Buri buremere bwerekana uburyo ikintu gihindura log-odd y'ibisubizo
  5. Nigute gusubira inyuma muburyo busanzwe bihindura ibishoboka mubirango byishuri?
    • Ikoresha urwego, akenshi 0.5
    • Buri gihe itora icyiciro cya mbere
    • Ugereranije abaturanyi
    • Ikoresha guhitamo
  6. Nigute gusubira inyuma kwa logistique bihuza imiyoboro yimitsi?
    • Ntaho bihuriye nabo
    • Urusobe rw'imitsi ya nyuma ya sigmoid cyangwa softmax urwego ni muburyo bwo gusubira inyuma
    • Isimbuza ibyiciro byose byihishe
    • Irashobora gukora kuri GPU gusa
  7. What does convexity tell us about the binary logistic-regression negative log-likelihood?
    • It guarantees exactly one finite solution for every dataset
    • It makes the model immune to mislabeled examples
    • A finite local minimum is global, but convexity alone does not guarantee uniqueness
    • It means the decision boundary must be nonlinear
  8. What can complete separation imply for unregularized logistic-regression maximum likelihood?
    • The model must misclassify every training example
    • The sigmoid stops mapping scores to probabilities
    • The optimizer automatically adds regularization
    • There may be no finite maximum-likelihood estimate because coefficients can grow without bound