Classification

Question 1

What type of classification is done in NLP?

Answer 1

sentiment analysis (also fake news detection)

Question 2

What problems does naive Bayes have?

Answer 2

• all features are equally important
• conditional independence assumption
• context not taken into account
• unknown words

Question 3

What problems does logistic regression solve that naive Bayes had?

Answer 3

• logistic regression assigns weights to features
• logistic regression tries to see which features are more important

The following stays unresolved:

• context not taken into account
• unknown words

Question 4

What problems does neural network solve that naive Bayes had?

Answer 4

fits more complex problems, whatever that means…

• NN assigns weights to features
• NN tries to see which features are more important

The following stays unresolved:

• context not taken into account
• unknown words

Question 5

What problems do CNNs solve that naive Bayes had?

Answer 5

CNNs capture context hierarchically.

• all features are equally important
• conditional independence assumption
• context not taken into account

The following stays unresolved:
- unknown words

Question 6

What problems do character level CNNs solve that naive Bayes had?

Answer 6

CNNs capture context hierarchically.

all of them yay!

Just an overview what the problems where:

• all features are equally important
• conditional independence assumption
• context not taken into account
• unknown words

Question 7

What problems do RNNs solve that naive Bayes had?

Answer 7

RNNs capture context sequentially.

all of them yay!

Just an overview what the problems where:

• all features are equally important
• conditional independence assumption
• context not taken into account
• unknown words

Question 8

Accuracy

Answer 8

accuracy = (true positive + true negative) / #data points

Question 9

F1 Score

Answer 9

f1_score = 2 * (precision*recall) / (precision + recall)