ranked_features.txt

Ranked features (1st draft):

higher the # of followers in context topic
golden ratio # of words in the question. 50 > count > 10
higher # of topics
sum of followers in topics
not anon
# of common nouns between question text and context topic
# of common nouns between question text and topics
Is it a yes or no question? (Is..will..can..do..does..are..)
What kind of question is it? (Who? What? Where? When? Why? How?)
no additional topics
question text count > 50
# of sentences
ends with a question mark
freq of punctuation
ratio of extraneous pronouns
ratio of verbs
ratio of adjectives
What is the average length of word?
if words are capitalized after a period
Does the question have a proper noun in it?
Does the question have a name in it?
Does the question have a name of someone famous in it? (list of celebrities)
Is the question related to technology?
Does the question contain the name of a poppin' startup?
Is the question related to science / mathematics?
Is the question related to pop culture?
Is the question related to sex / sexuality?
Is the question related to love / romance?
Is the question related to drugs?
Is the question related to money?
Does the question contain the word "best"?
Does the question contain the word "worst"?
Does the question contain the phrase "the most"
Does the question contain the phrase "the least"
Does the question contain a typo?
# of misspellings
Does the question address the reader directly?
Is the question a hypothetical / conditional question? (If ....would?)
Does the question contain words that do not exist in a colloquial english dictionary?
How many of the most commonly occurring words in the dataset (excluding articles and prepositions) does the question contain?