Quizz 06: Classification

1. Consider the feature extraction function we reviewed for POS tagging using a classifier:

   def features(sentence, index):
       """ sentence: [w1, w2, ...], index: the index of the word """
       return {
           'word': sentence[index],
           'is_first': index == 0,
           'is_last': index == len(sentence) - 1,
           'is_capitalized': sentence[index][0].upper() == sentence[index][0],
           'is_all_caps': sentence[index].upper() == sentence[index],
           'is_all_lower': sentence[index].lower() == sentence[index],
           'prefix-1': sentence[index][0],
           'prefix-2': sentence[index][:2],
           'prefix-3': sentence[index][:3],
           'suffix-1': sentence[index][-1],
           'suffix-2': sentence[index][-2:],
           'suffix-3': sentence[index][-3:],
           'prev_word': '' if index == 0 else sentence[index - 1],
           'next_word': '' if index == len(sentence) - 1 else sentence[index + 1],
           'has_hyphen': '-' in sentence[index],
           'is_numeric': sentence[index].isdigit(),
           'capitals_inside': sentence[index][1:].lower() != sentence[index][1:]
       }
   

   Indicate which of these features are of the following types:

   
   Lexical:
   Morphological:
   Syntactic:
   

2. Why is the classifier-based approach for POS superior to the backoff method of combining multiple taggers (for example, the unigram tagger with the affix-based tagger combined in a backoff manner)?

   
   
   
   
   

3. Explain the intuition behind TF*IDF word weighting for bag of words document encoding.

   
   
   
   
   
   
   
   

4. Why is it desirable to reduce the dimensions of the bag of words representation before applying a classifier?

   
   
   
   
   
   
   

5. Give three examples of dimensionality reduction used for bag of words representation