Features extracted (for both cb and non_cb):

1. get_no_of_tokens() : Number of words in the headline
2. get_avg_char_count() : Average number of characters in words (of headline)
3. sd_len(): Maximum length of syntactic dependency between governing & dependent words. If there exists one syntactic dependency of length 2 and two of length 4, the output vector for a given headline would be [0, 0, 1, 0, 2] i.e the value at the corresponding index would be 1 (number of such dependencies). The maximum number of tokens in any headline is 19, so a list with 19 elements has been created.
4. sub_count(): Get a list of all words used as subjects in a title