Algorithm 1: Feature extraction algorithm based on rules and opinion lexicon

Input: POS-tagged pre-processed sentence, feature-rule-lexicon, Opinion Lexicon

Output: Feature

Begin

1 for each POS-tagged-sentence “si” in POS-tagged sentences do

begin

2   while not (feature-rule-lexicon.eof)

begin

3     if (POS-tagged-sentence “si” matches rule in feature-rule-lexicon) then

begin

4     if (sentiment matches Opinion Lexicon) then

begin

5       Extract feature-sentiment N-gram along with rule # from feature-rule-lexicon

6      Store it in Feature-Opinion-Lexicon (L1)

End if

End If

End while

End for

End