All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
After carefully reviewing the revisions you have made in response to the reviewers' comments, I am pleased to inform you that your manuscript has been accepted for publication in PeerJ Computer Science.
Your efforts to address the reviewers’ suggestions have significantly improved the quality and clarity of the manuscript. The changes you implemented have successfully resolved the concerns raised, and the content now meets the high standards of the journal.
Thank you for your commitment to enhancing the paper. I look forward to seeing the final published version.
[# PeerJ Staff Note - this decision was reviewed and approved by Xiangjie Kong, a PeerJ Section Editor covering this Section #]
This work has been modified according to my comments, and the quality of this work can be accepted.
I don't have any comments.
No problems, it is good now.
It is good.
None.
Thanks to the authors for responding to the requested comments. However, the English in the newly updated texts contains some typos. I recommend reviewing the paper again before submitting it.
No more comments.
No more comments.
No more comments.
I have read the revised version of the paper carefully. The authors have taken all my suggestions into account. In my opinion the quality of paper has improved considerably. The writing quality, mathematical developments and numerical experiments appear to be good. Figures and tables are designed and organized effectively. I think that it provides a nice contribution to the scientific community. I have no further comments or suggestions to it.
In the light of the suggestions, the experimental results have proven the success of the proposed model, which has been expanded.
The authors compared their approach with the newly published models. The literature review was expanded with new references and the superiority of the proposed approach over other studies was emphasized.
Thank you for submitting your manuscript to PeerJ. After careful consideration of the reviewers' comments, we regret to inform you that major revisions are required before your paper can be considered further for publication.
The reviewers have raised several concerns regarding the methodology and experimental evaluation presented in your paper. These issues need to be thoroughly addressed to ensure the robustness and validity of your findings.
In addition, we recommend a comprehensive rereading of your manuscript to correct any typographical, grammatical, or formatting errors that may have been overlooked. Ensuring clarity and coherence in your writing will greatly enhance the readability and overall quality of your paper.
Please revise your manuscript accordingly and provide a detailed response to each of the reviewers' comments. We look forward to receiving your revised submission and appreciate your understanding and cooperation in improving the quality of your work.
[# PeerJ Staff Note: It is PeerJ policy that additional references suggested during the peer-review process should *only* be included if the authors are in agreement that they are relevant and useful #]
This work should introduce the research status of sentiment analysis deeply and comprehensively.
This work should carry out experiments on other emotion classification experiments, such as four categories, seven categories, etc.
Moreover, this work should compare with the newly published models.
This paper analyses the sentiment of Hajj pilgrims on social media to improve crowd management during Hajj by developing a CNN-LSTM based deep learning model. The study first collects and processes relevant datasets and then feeds them into the model for sentiment classification, and experimentally verifies the superiority of the model in terms of accuracy and effectiveness, and the results show that the model performs well in the sentiment analysis task and can provide valuable sentiment information for crowd management. But there are also problems that need to be modified.
The problems mainly include the following aspects:
1. In the abstract part, the article does not clearly describe the CNN-LSTM approach and I suggest adding a description of the approach.
2. In the abstract section, the sentence "while our proposed model on our novel dataset confirms that the proposed model outperforms other models by 92%." is not clear and is redundant.
3. In the introduction, while the article covers some research information on sentiment analysis, it does not go far enough in describing existing research and lacks detailed analyses of current techniques, methods and problems.
4. In the introduction, The article does not clearly describe the specific research question addressed by the article and lacks a detailed elaboration of the inadequacies of the existing methodology and a clear definition of the specific problem to be addressed by this study. It is recommended to list the questions in the form of 1.2.3 in order to more clearly express the research question in the research.
5. In the introduction, in the third contribution point, the article only states that comparisons with different models have been made, but it does not state the specific advantages of its own proposed model
6. In the experimental section, the article only compares the performance of the three classifications and does not compare other classifications such as the two classifications, the seven classifications etc.
The paper is written well. There are some issues I observed that must be resolved before publication as below:
1. Table 1: Summary of existing approaches must include batch dataset name and performance gained by the work.
2. Table 1: Summary of existing approaches must consist of a minimum of 20 existing works.
3. Table 5: The specifications of hardware and software must be more precise not such general.
4. The authors must specify the default values and optimized value of parameters in Table 6.
5. The result section is so weak. the novelty of the work must be specified.
6. The authors have failed to explain the choice of hyperparameters for optimization with CNN not justified or compared against other optimization techniques in section 5.2
7. The authors must include an ablation study to justify the performance of their work.
8. The validation of the results not found (K-fold approach)
9. There is no discussion of the computational complexity of this proposed model
10. The model interpretability section is superficial and does not provide actionable insights for real work scenarios.
11. There is insufficient explanation of the preprocessing steps and how they handle potential biases in the data in section 5.5.
12. The manuscript does not explore the limitations and potential risks of applying this hybrid model in the discussion section.
The novelty of the research is not clearly addressed. While there are many studies that have applied similar methods on different datasets, the authors need to explain their unique contribution to the field and identify the specific research gap they aim to fill.
The authors should provide a detailed explanation of the parameter values utilized in the experimental configurations. It is suggested that additional experiments be included to substantiate the rationale behind the chosen values.
There are many research works on sentiment analysis for pilgrims using the CNN-LSTM method that are not included in the related work. Strengthen the literature review by including more recent studies such as:
Elgamal, M., 2016. Sentiment Analysis Methodology of Twitter Data with an application on Hajj season. International Journal of Engineering Research & Science (IJOER), 2(1), pp.82-87.
Gutub, A., Shambour, M.K. and Abu-Hashem, M.A., 2023. Coronavirus impact on human feelings during 2021 Hajj season via deep learning critical Twitter analysis. Journal of Engineering Research, 11(1), p.100001.
Shambour, M.K., 2022. Analyzing perceptions of a global event using CNN-LSTM deep learning approach: the case of Hajj 1442 (2021). PeerJ Computer Science, 8, p.e1087.
Khan, M.A. and AlGhamdi, M., 2024. A customized deep learning-based framework for classification and analysis of social media posts to enhance the Hajj and Umrah services. Expert Systems with Applications, 238, p.122204.
Aldhubaib, H.A., 2020. Impressions of the community of Makkah on the Hajj in the light of Covid-19 pandemic: quantitative and AI-based sentiment analyses. Journal of King Abdulaziz University: Eng Sci, 32(1).
Albahar, M., Gazzawe, F., Thanoon, M. and Albahr, A., 2023. Exploring Hajj pilgrim satisfaction with hospitality services through expectation-confirmation theory and deep learning. Heliyon, 9(11).
Alghamdi, H.M., 2024. Unveiling Sentiments: A Comprehensive Analysis of Arabic Hajj-Related Tweets from 2017–2022 Utilizing Advanced AI Models. Big Data and Cognitive Computing, 8(1), p.5.
Gandhi, U.D., Malarvizhi Kumar, P., Chandra Babu, G. and Karthick, G., 2021. Sentiment analysis on twitter data by using convolutional neural network (CNN) and long short term memory (LSTM). Wireless Personal Communications, pp.1-10.
Alayba, A.M., Palade, V., England, M. and Iqbal, R., 2018. A combined CNN and LSTM model for Arabic sentiment analysis. In Machine Learning and Knowledge Extraction: Second IFIP TC 5, TC 8/WG 8.4, 8.9, TC 12/WG 12.9 International Cross-Domain Conference, 2018, Proceedings 2 (pp. 179-191).
Ombabi, A.H., Ouarda, W. and Alimi, A.M., 2020. Deep learning CNN–LSTM framework for Arabic sentiment analysis using textual information shared in social networks. Social Network Analysis and Mining, 10, pp.1-13.
**PeerJ Staff Note:** It is PeerJ policy that additional references suggested during the peer-review process should only be included if the authors are in agreement that they are relevant and useful.
Provide a deeper analysis and interpretation of the results obtained. Discuss not only the accuracy rates but also the implications of these results on the field of sentiment analysis and crowd management.
All charts, including diagrams or flowcharts, need to be redrawn to improve the paper's presentation.
The authors focus on analyzing a real case study related to preprepared food at Hajj 1442. However, the dataset was collected using many hashtags and keywords relevant to catering and the costs of Hajj campaigns, not only "catering".
Ensure the language is clear, concise, and free of ambiguities. Proofread the paper for grammar and spelling errors. Consider restructuring ambiguous sentences and correcting typographical mistakes for better readability, such as:
Line 136: "research, Alasmari et al. proposed"
Line 196: "data. Then, the biggest value is determined in each feature map by the max pooling layer."
Line 237: "Must be annotated (labeled) before the training stage."
Line 286: "(www) are removed from the tweets to improve their content. We use the re library, ...."
Line 299: "such as 6F', N.F', '0', 'G'."
Line 356: "get matrices of vectors representing texts. The dot product of vectors might produce"
Line 357: "when they belong to the same category by One-Hot Encoder, while if they find similarities"
Line 410: "determined via the sigmoid function, which takes the outputs of the last LSTM unit (ht21) at"
Line 425: "where Ct21"
Line 454: "Our novel dataset,"
Line 554: "Figure 15 shows the results of our proposed method performing better on our dataset with an F1-score."
avoid mentioning "novel dataset"
This manuscript discusses a study examining crowd management and pilgrims’ experiences during Hajj. Saudi authorities use modern technologies to ensure the comfort of pilgrims. The authors emphasize that the emotions and experiences of pilgrims through social media are not well known. This study used Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) algorithms to analyze the emotions of pilgrims. The model was tested on a case study of pre-prepared meals during Hajj 1442, compared with other machine learning algorithms, and was observed to give the best results with a success rate of 92%.
The subject discussed in the study is an original field that has yet to be studied much. However, the method proposed in the study is a hybrid of CNN and LSTM methods, which have also been used in other studies. The difference between the proposed new model and other models should be emphasized. The originality of the proposed model can be increased by increasing the convolution layers, which have been proven decisive in capturing local features.
The one-hot-encoding word embedding technique's successful performance in the proposed model has fallen far behind the current technological word embedding vectors (word2Vec, GloVe, fastText). Is there a particular reason for choosing one-hot encoding in the study? Otherwise, using other word embedding techniques, the proposed model should not be revised again.
A more thorough explanation of the suggested approach and a thorough examination of its computational complexity are necessary.
WordClouds should be used when analyzing sentiment distributions. Statistical data does not include the reasons for the pilgrimage thoughts in the content; it only provides information about the statistical majority.
The authors should also contrast their work with previously published works.
The introduction section requires more details about the originality and novelty of the approach proposed against what has been developed before.
Although I think the subject discussed in the study is interesting, I believe that the experimental results are not satisfactory.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.