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.
Revision of review comments is done comprehensively
[# PeerJ Staff Note - this decision was reviewed and approved by Xiangjie Kong, a PeerJ Section Editor covering this Section #]
No comment
No comment
No comment
Thanks for Incorporating all the comments
I think the paper has been revised in a good manner
I think the paper has been revised in a good manner
I think the paper has been revised in a good manner
Please follow review comments
[# PeerJ Staff Note: The review process has identified that the English language must be improved. PeerJ can provide language editing services if you wish - please contact us at [email protected] for pricing (be sure to provide your manuscript number and title). Your revision deadline is always extended while you undergo language editing. #]
Clarity of Methodology: The manuscript lacks clarity in explaining the methodology employed for evaluating the hybrid architecture. Please provide more detailed descriptions of the experimental ko, including parameters, metrics, and datasets used.
Quality of Writing: Pay attention to the quality of writing, including grammar, punctuation, and clarity of expression. Ensure that the manuscript is well-written and free of typographical errors.
Statistical Analysis: The results presented lack statistical analysis, making it challenging to assess the significance of the findings. Please include statistical tests to validate the observed differences and improvements over existing approaches.
Discussion on Limitations: While the limitations of the study are briefly mentioned, a more comprehensive discussion is needed. Please elaborate on potential weaknesses in the methodology, assumptions made, and uncertainties in the results.
Validation of Simulation Environment: Please provide validation of the simulation environment against real-world data to ensure the relevance and reliability of the findings.
Discussion on Scalability: Please provide a detailed discussion on how the architecture scales with increasing numbers of IoT devices and data volume.
Validation against Real-World Scenarios: While simulated environments provide insights, validation against real-world scenarios is essential. Consider conducting pilot tests or case studies in real-world settings to validate the performance of the architecture.
Comparative Analysis Clarity: The comparative analysis with existing approaches lacks clarity in explaining the criteria used for comparison. Please clearly define the evaluation criteria and justify why the proposed architecture outperforms alternatives in each aspect.
Reproducibility: Ensure that all experimental procedures and configurations are thoroughly documented to facilitate reproducibility by other researchers. Include details on software versions, hardware configurations, and any custom code or algorithms used.
Ethical Considerations: Discuss any ethical considerations associated with the proposed architecture, particularly regarding data privacy, consent, and potential societal impacts.
N/A
This paper introduces a novel technique called Word Embedding LDA (WELDA) for improving topic recognition in news articles by integrating topic models with word embeddings. The primary motivation is to address the lack of standardized methods for combining these models to enhance text representation and achieve higher accuracy in downstream natural language processing tasks.
Key Contributions:
1.Novel Topic Model - WELDA:
Integrates Latent Dirichlet Allocation (LDA) with Word2Vec embeddings.
Uses a modified Gibbs Sampling algorithm to enhance topic recognition.
Achieves significant improvements in accuracy for news topic recognition, with results showing 88% accuracy on the 20NewsGroup dataset and 97% on the BBC News dataset.
2. Framework for News Text Data:
Comprised of three core modules: Preprocessing, Feature Extraction, and Topic Recognition.
Utilizes classifiers such as Random Forest and Logistic Regression to leverage the enhanced text representations.
3. Comprehensive Evaluation:
The method was rigorously tested on benchmark datasets (20NewsGroup and BBC News) and compared against several baseline techniques.
Demonstrated superior performance in terms of accuracy and topic recognition metrics.
Datasets: 20NewsGroup and BBC News.
Methods: Comparisons were made against TF-IDF, LDA, Word2Vec, Glove, and several hybrid models (SGL, CGL, SWL, CWL).
Metrics: Evaluated using Micro-F1 score and classification accuracy.
Literature Review:
Current Strengths: The paper presents a thorough literature review of existing methods in topic modeling and word embeddings.
Suggestions for Improvement:
Deeper Analysis: Include a more detailed comparison of the limitations of existing methods and how WELDA specifically addresses these gaps. For instance, a tabular comparison highlighting key differences and improvements could add clarity.
Broader Scope: Consider discussing more recent advancements in neural topic models and their relevance to WELDA.
Methodology:
Current Strengths: The methodology is well-structured and clearly explained, with a detailed description of the WELDA model and the modified Gibbs Sampling algorithm.
Suggestions for Improvement:
Complexity Analysis: Include a discussion on the computational complexity of the proposed method compared to baseline models.
Current Strengths: Results are well-presented with tables and graphs comparing the performance of WELDA against other models.
Suggestions for Improvement:
Error Analysis: Provide a more detailed error analysis to understand where and why the model fails in certain cases. This could involve a qualitative examination of misclassified examples.
Ablation Studies: Conduct ablation studies to isolate the impact of each component of the WELDA model (e.g., the contribution of word embeddings versus topic modeling).
5. Conclusion and Future Work:
Current Strengths: The conclusion summarizes the contributions and suggests future research directions.
Suggestions for Improvement:
Implementation Details: Share more insights on the implementation challenges and potential solutions, which can guide future researchers.
Broader Implications: Discuss the broader implications of the findings in real-world applications, such as media monitoring and automated content analysis.
Extended Experiments: Suggest extending the experiments to include streaming data scenarios to test the model’s adaptability to real-time applications.
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.