Affective-Knowledge-Enhanced Graph Convolutional Networks for Aspect-Based Sentiment Analysis with Multi-Head Attention
Round 1
Reviewer 1 Report
The paper proposes a novel Graph Neural Network model, MHAKE-GCN, for ABSA that incorporates external sentiment knowledge into the graph convolutional neural network and fully extracts semantic and syntactic information from the sentence using multi-head attention.
The model aims to improve the learning of sentiment expressions related to specific aspects by adding weight to sentiment words associated with aspect words.
The model was evaluated on four publicly benchmark datasets and compared against twelve other methods. The experimental results demonstrated the effectiveness of the proposed model for the ABSA task.
The paper provides a detailed introduction to the ABSA task and its challenges and discusses related work. The proposed model and its contributions are clearly presented and well-explained.
The references are adequate and up to date.
Overall, the paper appears to be well-written and makes a valuable contribution to the field of sentiment analysis.
Author Response
Thanks very much for taking the time to review this manuscript. We appreciate all your comments!
Reviewer 2 Report
This paper is related to an interesting topic, namely Affective Knowledge-Enhanced Graph Convolutional Networks for Aspect-Based Sentiment Analysis with Multi-Head Attention for Applied Sciences
However, the manuscript needs to be further developed in order to meet the expected academic requirements.
1) The related work section is weak. Add more details and studies in Related Work section.
2) Add the discussion section and should be strengthened with more details.
3) Add the Implications section
4) The conclusions are concise. An in-depth reflection about the results should be discussed.
Author Response
Point 1: The related work section is weak. Add more details and studies in Related Work section.
Response 1: We are grateful for the suggestion. We have added more details and studies in related work section.
Point 2: Add the discussion section and should be strengthened with more details.
Response 2: We are grateful for the suggestion. We have added the discussion section and we have discussed our research further.
Point 3: Add the Implications section
Response 3: We are grateful for the suggestion. We have added the Implications section.
Point 4: The conclusions are concise. An in-depth reflection about the results should be discussed.
Response 4: We are grateful for the suggestion. We have revised our conclusions and have summarised our research in more depth.
Reviewer 3 Report
What kind of data set is created. Any access on Github
Sec 4.2 what is the reason to explain models again here. ir it should be in a related work os system overview section.
Author Response
Point 1: What kind of data set is created. Any access on Github
Response 1: We are grateful for the suggestion. A sample of these open datasets consists of a raw text, an aspect of the text and its corresponding sentiment polarity, and the datasets were obtained from the following Github address:
https://github.com/yangheng95/ABSADatasets/tree/v2.0/datasets
Point 2: Sec 4.2 what is the reason to explain models again here. ir it should be in a related work os system overview section.
Response 2: We are grateful for the suggestion. Sec 4.2 is not explaining the model, but illustrating the experimental parameter settings of the model.