Aspect-Based Sentiment Analysis Using Aspect Map

Round 1

Reviewer 1 Report

Aspect-based sentiment analysis is a very important issue in current computer science. The authors propose a novel CNN-based aspect-level sentiment classification model, which  consists of two CNNs. The paper is well written and easy to follow. The approach is interesting and relevant, so this reviewer would argue for accepting it, and has only two comments for improving the paper: somewhere in the introduction the authors should mention that text-preprocessing is still always an important aspect and should not be underestimated – even in such approaches, see e.g. 1.  Petz, G. et al. On text preprocessing for opinion mining outside of laboratory environments. In Active media technology, lecture notes in computer science, lncs 7669, Huang, R.; Ghorbani, A.; Pasi, G.; Yamaguchi, T.; Yen, N.; Jin, B., Eds. Springer: Berlin Heidelberg, 2012; pp 618-629. 10.1007/978-3-642-35236-2_62

One of the big drawbacks of this approach is inherent in CNNs and is in difficult re-traceability, re-enactivity, explainability, so this is an open problem for future research which can be mentioned in the future work section, here a very current pointer: 2.       Holzinger, A. From machine learning to explainable ai. In 2018 world symposium on digital intelligence for systems and machines (ieee disa), 2018; pp 55-66. 10.1109/DISA.2018.8490530

Overall, this is a very pleasant paper and this reviewer would recommend acceptance.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

This paper proposes a novel CNN-based aspect-level sentiment classification model. The proposed model has been applied on SemEval 2016 Task 5 dataset and is compared with several baseline models. According to the experimental results, the proposed model does not only outperform the baseline models but also shows state-of-the-art performance for the dataset. The paper is well written and technically sound.

My suggestions are the following:

- I suggest to improve the introduction of the paper. A more detailed description of the state of the art is needed. 

- The description of the proposed approach is too difficult to read. I suggest to simplify the section.

- A more detailed discussion about the obtained results is needed.

Author Response

Please see the attachment.

Author Response File: Author Response.docx