Dear Researcher

you can use different tools for fuzzification.you can check my publications in which i have done fuzzification taking samples.

please refer to the attachments which will help you.

regards

http://www.mathworks.com/help/pdf_doc/fuzzy/fuzzy.pdf

http://jfuzzylogic.sourceforge.net/html/manual.html

First, You have to come up with with the linguistic  variable and the level. What are you going to measure? temperature (hot, cold, very hot, very cold), customer satisfaction (high, low, very low, etc)

Then, define the corresponding scale of crisp value for each linguistic variable.

Third, determine the fuzzy number type whether Triangular or trapezoidal. For TFN (Triangular Fuzzy Number) you have to set the alpha cut  0< alpha < n (n could be any positive number). for example crisp value 3 would be TFN by form of  (l, m, u) 3-alpha, 3, 3+alpha

Averaging data won't give same result. For defuzzification you can use several methods which are centre of gravity, first of maxima, last of maxima, weight of maxima etc.. and for those methods will provide different result. You have to adopt the appropriate method representing the shape of Fuzzy Number.