Kedis-ka-duwanaanta Ganacsiga

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Qaabka aad u fudud in lagu qabto qaabka hoose ee xogta ayaa ugu horrayn la il daran:
   • Kala duwanaansho sare
   • Eexda sare
   • Khalad hoose oo aan la xakamayn karin
   • Qalab xad dhaaf ah
2. Xaaladdee ayaa calaamad u ah kala duwanaanshiyaha sare (ku-fiicnaanta xad dhaafka ah)?
   • Tababarka iyo khaladaadka imtixaanku labaduba waa kuwo sarreeya oo la mid ah
   • Qaabka ayaa iska indha tiraya sifooyinka badankooda
   • Ku darida joogtaynta dheeraadka ah waxay kordhisaa qaladka tababarka ee ku wajahan khaladaadka imtixaanka
   • Khaladka tababarku aad buu u hooseeyaa laakiin khaladka imtixaanku aad buu u sarreeyaa
3. In the classic bias-variance picture described in the guide, which tradeoff can accompany increased model flexibility?
   • Eexda way yaraataa kala duwanaanshuhuna wuu kordhaa
   • Labaduba way kordhaan
   • Labaduba way yaraadaan
   • Eexda way korodhaa, kala duwanaanshuhuna wuu yaraadaa
4. Ku darida L2 (jidh) joogtaynta moodeel guud ahaan:
   • Waxay yaraynaysaa eexda iyadoo ku kacaysa kala duwanaansho badan
   • Waxay tirtirtaa khaladka aan la joojin karin
   • Waxay yaraynaysaa kala duwanaanshaha qiimaha eex yar oo dheeraad ah
   • Wax raad ah kuma laha eexda ama kala duwanaanshiyaha
5. Khaladka aan la dhayalsan karin ee kala-duwanaanshaha eexdu waxay u taagan tahay:
   • Khalad ka dhashay doorashada algorithm khaldan
   • Khalad ka yimid buuqa ku jira xogta oo aan noocna ka saari karin
   • Khalad ku baaba'a xogta tababarka dheeraadka ah
   • Khaladka waxaa keenay kaliya kala duwanaansho sare
6. Dhacdadii 'laba-jileec' ee lagu arkay barashada qoto dheer waxay soo jeedinaysaa in:
   • Cilada tijaabadu mar kale way yarayn kartaa marka moodellada ka koraan meel ka fog barta xogta tababarka
   • Cilada imtixaanku waxay mar walba raacdaa qaabka U-fudud
   • Kala duwanaanshuhu waligii dhib uma laha shabakadaha neerfaha
   • Eexda had iyo jeer waa eber moodooyinka qoto dheer
7. In a squared-error decomposition, squared bias is 2, variance is 3, and irreducible noise is 1. What is the total expected error?
   • 14
   • 5
   • 3
   • 6
8. Why can averaging many de-correlated trees help in the random-forest example?
   • It removes all noise from the data-generating process
   • It guarantees zero test error
   • It can reduce the variance of the combined prediction
   • It converts every tree into a linear model