Text Data Augmentation for the Korean Language
Round 1
Reviewer 1 Report
There are some points that need to be further clarified:
- Rewrite the Abstract. The experimental result and the performance of your work must be highlighted in the Abstract.
- The introduction of the method is not clear enough. It is necessary to supplement the description of the mathematical model.
- Thereby, more experiments, such as the K-fold cross-validation test, should be implemented to reveal the advantages of the proposed approach.
- Some remarks on the main results would be necessary and helpful.
Author Response
1. Rewrite the Abstract. The experimental result and the performance of your work must be highlighted in the Abstract.
Thank you for these observations. We have rewritten the abstract to better differentiate among the objectives and edited it so that the performance of our work can be addressed in the Abstract. (lines 16-24)
2. The introduction of the method is not clear enough. It is necessary to supplement the description of the mathematical model.
We agree with the Reviewer's suggestion and have added descriptions of the mathematical model. The modification can be found in lines (172-176, 198, 203, 208, 212, 233-235).
3. Thereby, more experiments, such as the K-fold cross-validation test, should be implemented to reveal the advantages of the proposed approach.
We agree with the Reviewer's consideration. As mentioned in the Experimental results part, to ensure the randomness and confidence of our results, we have experimented with cross-validation, as shown in Table 3 and Table 4. We describe details of our settings at lines 290-304.
4. Some remarks on the main results would be necessary and helpful.
We agree with the Reviewer's suggestions. We have provided some remarks in the main result, at lines 341-352
Reviewer 2 Report
The article, being a novel contribution for the data augmentation in the case of the Korean Language does reflect a significant effort of authors. In particular, the richness of the Datasets the authors used in the experimental section is impressive.
The text is well structured and written. Notwithstanding the minor omissions in several points that follow below, the text requires no further particular effort from the authors.
The introduction can be organized as a unique section without sub-sections (1.1, 1.2 etc.). The passages from one paragraph to another are quite distinct, so there is no necessity to divide them into separate sub-sections.
On line 277, please control whether you've meant "per one sample" instead.
Line 296, correct the "forssssss".
Finally, the Reviewer considers the Conclusion section a bit concise. Try to add some details related to the experimental work.
Author Response
1. The introduction can be organized as a unique section without sub-sections (1.1, 1.2 etc.). The passages from one paragraph to another are quite distinct, so there is no necessity to divide them into separate sub-sections.
Thank you for these observations. We have rearranged the introduction and removed sub-sections.
2. On line 277, please control whether you've meant "per one sample" instead.
Thank you for pointing out our typo. We have been replaced "once" with "one".
3. Line 296, correct the "forssssss".
Thank you for pointing out our typo. We have been replaced "forssssss" with "for".
4. Finally, the Reviewer considers the Conclusion section a bit concise. Try to add some details related to the experimental work.
We agree with the Reviewer's suggestion and have added details into the Conclusion. The modification can be found in lines (358-363).
