How can scikit-fuzzy is used in document summarisation

[Dannykl Ng]

Dear all

I am working on query-based document summarization project that involves few steps including fuzzy logic. I would like you to help me with how can I achieve the fuzzy logic part using scikit-fuzzy. Please note, I am new to scikit-fuzzy or fuzzy logic implementation. I have basic knowledge on how fuzzy logic works.

Here is the problem:

I have around 4000 sentences and features associated with the sentences. The next line shows the example format 

list = [("The pun - ancient artform , or the lowest form of wit ?"), (0.14256387543382126,), (0,), (5,), (12,), (0,)]

Explanation about above line [ (the row sentence), (the similarity value this sentence with query), (number of proper noun), (number of noun), (sentence length) and (number of digital values)]

I would like to determine how good the sentence is based on the similarity value, number of proper noun, number of noun, length of the sentence and number of digital values using the scikit-fuzzy.

For instance 

IF the similarity value >= 0.9 AND the number of proper noun >= 2 AND number of digital value >=1 THEN the sentence is very important

IF the similarity value >= 0.6 AND the number of proper noun >= 1 AND number of digital value >=1 THEN the sentence is important

IF the similarity value <= 0.1 AND the number of proper noun = 0 AND number of digital value >=0 THEN the sentence is very poor 

categories are very important, important, medium, poor, very poor

Can you help me to implement this in scikit-fuzzy? 

Apologies in advance if i did not make it clear and precise. Also, reply to me if you require more information. 

Regards,

Daniel