Input: parts of opinion from college and university students after preprocessing. The total feature used K that illustrates the vector length of every feature.
Output: feature Vector
Step 1: on the given opinion, estimation of frequency for every opinion and the outcome is stored by using hash table where the opinion index is given as key and value is the term frequency. The opinion index is estimated as follows: . Hash table is utilised to additional vector which depicts the CNN vector of opinion .
Step 2: on every opinion, estimate the vector frequency of the opinion for every opinion o as follows:
for in this opinion O:
for index in this oj
of
end
end
Step 3: on the opinion given by students, aggregate frequency for every vector is obtained from all the opinion obtained by students and values to the corresponding opinion are accumulated to get the vector , including to every node in the network.
Step 4: on every opinion, estimate the CNN feature vector for every opinion using (3).
