Combining Natural Language Annotation with Transformer for Thai Sentence Classification
Saranlita Chotirat, Phayung Meesad · 2020
Sentence classification is one part of the NLP tasks that have a wide range of applications including question answering, sentiment analysis, and much more. The Challenge of sentence classification is more ambiguous because they might have not enough contextual information. While this is not a simple task, it is becoming increasingly important to many applications. In this study, we propose to classify sentences to a wh-question category based on combined Natural Language Processing annotation (Part-of-Speech tags and Named Entity Recognition) and attention transformer mechanism. The experimental results on both interrogative sentences and simple sentence datasets showed that the proposed model improves the performance of sentence classification to the wh-question category with promising accuracy.