Research Article
Bullet Subtitle Sentiment Classification Based on Affective Computing and Ensemble Learning
Input: bullet subtitle , the Chinese Bullet Subtitle Sentiment Lexicon | Output: the sentiment value of bullet subtitle under sentiment category | Step 1: extract all sentiment words belonging to sentiment category in ; initialize , , | Step 2: find all the degree adverbs and the number of negatives between the sentiment word and the previous sentiment word | Step 3: find all the sentiment punctuation between the sentiment word and the next sentiment word | Step 4: use equation (6) to obtain the sentiment reversal variable | Step 5: use equation (5) to calculate the text sentiment value of the current sentiment word | Step 6: when is greater than or equal to , execute Step7; otherwise, execute Step2, , | Step 7: calculate the gain intensity of using the corresponding dimensional features of and equation (2); | Step 8: output and the algorithm ends. |
|