WorksheetsNLP3
Total questions: 9
Worksheet time: 5mins
Name
Class
Date
1.
1. What is the main focus when evaluating a binary classification model's performance?
a)
A) True Negatives (TN)
b)
B) False Positives (FP)
c)
C) False Negatives (FN)
d)
D) True Positives (TP)
2.
Which metric considers the ratio of correctly classified positives to all predicted positives?
a)
A) Precision
b)
B) Recall
c)
C) Accuracy
d)
D) F1 Score
3.
.In multiclass classification, how is the "positive" class defined for each class?
a)
A) As the class with the highest frequency
b)
B) As the class with the lowest frequency
c)
C) Arbitrarily for any class we are focusing on
d)
D) As any class except the current one
4.
What is the formula for calculating the F1 Score, a combined metric for precision and recall?
a)
A) F1 Score = Precision / Recall
b)
B) F1 Score = 2 * (Precision * Recall) / (Precision + Recall)
c)
C) F1 Score = Precision - Recall
d)
D) F1 Score = Precision * Recall
5.
When evaluating a binary classification model, if the Accuracy is 90%, and 90% of the actual data are positive classes?
a)
This is a very good model
b)
This is a very bad model
c)
Accuracy is not the proper metric in this case
d)
This is a sign of overfitting
6.
When evaluating a binary classification model, and all positive instances are correctly classified . Then
a)
Precision = 0
b)
Recall = 0
c)
Precision = 1
d)
Recall = 1
7.
When evaluating a binary classification model, and all instances are classified as negative. Then
a)
Precision=0 and Recall=0
b)
Precision=1 and Recall =1
c)
Recall = Precision=Accuracy
d)
Recall=1
8.
Which of the following is NOT a common vectorization technique used in text classification?
a)
Bag of Words (BoW)
b)
Term Frequency-Inverse Document Frequency (TF-IDF)
c)
Word Embeddings
d)
Text preprocessing
9.
When splitting a dataset into training and validation sets, what is the main reason for using stratified sampling?
a)
A) To ensure equal class distribution in both sets
b)
B) To randomize the data for better model performance
c)
C) stratified sampling is always preferred in Regression and Classification
d)
D) To simplify the text classification process
100 %
