ML_workshop_DAY_4_03/08/23

Testing phase comes after which phase?

• Data Collection phase

• Optimization phase

• Training Phase

• Deployment Phase

What percentage of the dataset is usually reserved for normal testing?

• 5-10 %

• 30-50%

• 100%

• 10-30%

Why is Cross Validation better ?

(only two options are correct)

• Much more robust

• Testing is done across the

  whole dataset fully

• It saves time

• It gives you high accuracy

Is precision and recall always preferred over accuracy?

• Sometimes

• Always

• You never know for sure

• No idea

Recall depends on ?

• True Positive and

  False Positive

• True Negative and

  False Negative

• True Positive and

  False Negative

• True Positive and

  True Negative

Characteristics of Overfitting model ?

• High Training Error

  and

  High Testing Error

• Low Training Error

  and

  High Testing Error

• Low Training Error

  and

  Low Testing Error

• High Training Error

  and

  Low Testing Error

What are Hyperparameters?

• Values/variables which

  influence the training process

• Values which are learnt

  from the dataset

• Parameters which are hyper

• hyper values which are

  parameters

What happens after a good hyper parameter optimization ?

• Best Hyper Parameter

  values are chosen and

  trained

• Nothing, training is finished

• Restart the whole

  training process

• Pickling the base model

How was today's session?