Journal
COGNITIVE COMPUTATION
Volume 13, Issue 5, Pages 1114-1127Publisher
SPRINGER
DOI: 10.1007/s12559-021-09855-4
Keywords
Aspect based; Sentiment analysis; Machine learning; Cognitive computing
Funding
- National Key R&D Program of China [2020YFB1006002]
Ask authors/readers for more resources
By proposing a deep-level semi-self-help sentiment annotation system based on BERT and a novel classification model, the accuracy of ABSA tasks can be improved, space-time complexity can be reduced, and the amount of data annotation engineering required can be significantly decreased.
Aspect-based sentiment analysis (ABSA) can help consumers provide clear and objective sentiment recommendations through massive quantities of data and is conducive to overcoming ambiguous human weaknesses in subjective judgments. However, the robustness and accuracy of existing sentiment analysis methods must still be improved. We first propose a deep-level semiself-help sentiment annotation system based on the bidirectional encoder representation from transformers (BERT) weakly supervised classifier to address this problem. Fine-grained annotation of restaurant reviews under 18 latitudes solves the problems of insufficient data and low label accuracy. On this basis, bagging traditional machine learning algorithms and annotation systems, a novel classification model for specific aspects is proposed to explore consumer behavior preferences, real consumer feelings, and whether they are willing to consume again. The proposed approach can effectively improve the accuracy of the ABSA tasks and reduce the space-time complexity. Moreover, the proposed model can significantly reduce the quantity of data annotation engineering required.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available