Key points are not available for this paper at this time.
This study examines how the Support Vector Machine (SVM) combined with natural language processing techniques can be used to identify product features from free-text customer reviews. To verify the validity of the proposed approach, 22,157 restaurant reviews are collected and 3,701 sentences are randomly selected and manually annotated. The experiment results show that the average precision and recall are both higher than those of the Maximum Entropy (ME) based approach.
Chuanming Yu (Thu,) studied this question.