Aiṣedeede Kilasi ati Idanwo Tuntun

Awọn idahun ati awọn alaye han bi o ṣe n mu ibeere naa.

1. Kini idi ti išedede itele jẹ metiriki ti ko dara fun iṣoro isọdi ti ko ni iwọntunwọnsi?
   • Awoṣe kan le ṣe Dimegilio giga pupọ nikan nipa sisọ asọtẹlẹ kilasi to poju nigbagbogbo
   • Yiye ko le ṣe iṣiro nigbati awọn kilasi ko ba dọgbadọgba
   • Yiye nikan ṣiṣẹ fun ipadasẹhin
   • O nilo data iwọntunwọnsi pipe lati ṣe asọye
2. Kini SMOTE ṣe?
   • Npa awọn ayẹwo-kilasi poju rẹ lọ laileto
   • Ṣe ipilẹṣẹ awọn ayẹwo kekere sintetiki nipasẹ sisọ laarin awọn aladugbo
   • Ṣe alekun oṣuwọn ẹkọ fun awọn kilasi toje
   • Ṣe iyipada iṣoro isọdi si ipadasẹhin
3. Ọna wo ni o ṣe aiṣedeede LAYI iyipada nọmba awọn apẹẹrẹ ikẹkọ?
   • Iṣeduro apọju
   • Undersampling awọn poju
   • Iwọn iwuwo kilasi ni iṣẹ isonu
   • SMOTE
4. Kini eewu bọtini kan ti lilo iwọn apọju ṣaaju pipin data rẹ sinu ọkọ oju irin ati awọn eto idanwo?
   • O mu ki ikẹkọ losokepupo
   • O npa kilasi to poju kuro
   • O nigbagbogbo fa underfitting
   • Àdáwòkọ tabi sintetiki to nkan ojuami le jo sinu awọn igbeyewo ṣeto, infating ikun
5. Kí ni akọkọ downside ti ID undersampling?
   • O danu data kilasi-pupọ ti o le wulo
   • O ṣẹda data iro ti o le jẹ otitọ
   • O ṣiṣẹ nikan fun data aworan
   • O nigbagbogbo mu akoko ikẹkọ pọ si
6. Metiriki wo ni o yẹ julọ fun iṣiro awoṣe kan lori data to toje?
   • Ìwò išedede
   • Precision-Recall AUC tabi F1 Dimegilio
   • Aṣiṣe onigun mẹrin tumọ si
   • Nọmba ti awọn akoko ikẹkọ
7. What does adjusting the decision threshold after training change?
   • The number of features in the model
   • The class proportions in the training data
   • The trade-off between which examples are labeled positive and negative without changing the training examples
   • The model’s learned weights automatically
8. Which metric does the guide name as an additional option for rare-positive classification?
   • Mean squared error only
   • BLEU score
   • Silhouette coefficient
   • Matthews Correlation Coefficient (MCC)