Experimental Setup

Question 1

The Training Dataset is used to ___ the parameters of the model

Answer 1

fit

Question 2

The Validation Dataset is used to ___ the model with the parameters selected during the ___ and to adjust ___

Answer 2

test
training phase
hyper-parameters

Question 3

The Test Dataset is used to provide an ___ of a final model fit on the training dataset

Answer 3

unbiased evaluation

Question 4

Overfitting happens when the techniques used tend to exploit relations in the training data that ___

Answer 4

do not generalize to the rest of the data

Question 5

The K-fold Cross-Validation works by ___ the data in K sets and one by one a set is selected as ___

Answer 5

splitting

test set

Question 6

Error Measurement is the process of computing the error for the ___ and the ___

Answer 6

training set

test set

Question 7

Balanced dataset are datasets where all classes are ___ while in imbalanced datasets some classes are ___

Answer 7

equally represented

more prevalent than others

Question 8

We can ___ the training set but never the validation and tes sets

Answer 8

balance

Question 9

A baseline is the ___ of a ___

Answer 9

result

very basic model

Question 10

Cross-Validation is mostly used to get a more certain value of ___ for our model or to helps us ___ it better or select ___

Answer 10

performance
tune
another one

Question 11

How to calculate the F1-score?

Answer 11

2* (Precision * Recall) / (Precision + Recall)