Optimising Health Emergency Resource Management from Multi-Model Databases
Round 1
Reviewer 1 Report
The emengercy accident detection topic is important and timely. The authors try to do so and it should be commended. The research is carried out correctly, but no reliable conclusions are drawn. You should better describe Figure 6 (graph), e.g. add more figs and bring closer important fragments (differences and similarities). Then, more should be written in the summary (you are currently only providing two values). You should also compare your results to other works.
Minor comments:
- The article contains two drawings numbered 6.
- Figures 5 and 6 (correct number 7) are big and don't add anything important. You should add more information.
- Caption of all drawings should be contain sources. The description of Figure 4 is too general. All pictures are mis-centered.
- Lines 178, 274-280 are badly formatted.
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Multimodal in data science usually means multiple media like images, text, audio etc. In that sense, the title is a bit misleading.
A significant portion of the paper is unnecessarily devoted to the tables listing various attributes, some of which are not relevant to the target variable.
Figures need to be explained better.
GLM as the name and the reference cited indicate, contain multiple algorithms. When the authors say "GLM has an error of 9%," the reader becomes suspicious if the methodology is even sound.
The paper lacks enough technical depth and novelty needed of a research publication.
Please fix language issues: "There are interesting researchs", "More over, while data are beeing," etc
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
This manuscript intends to make contributions to the health sector by optimizing health emergency resource management. To achieve the optimization, the multimodal databases were
The purpose is unclear at the beginning of the abstract. The meaning of "A powerful database" is unclear to me. Can it be referred to as the "multimodal databases"? Or, due to the last eight years of collection, the database is qualified to be regarded as powerful but still not sure how the powerful database can be related to a multimodal one? The abstract does not provide clear methodologies, except for the success rate. Whether the higher the success rate is the optimization that this study intends to achieve its purpose.
In the Introduction section, "the predictive model... the effectiveness of these predictive algorithms..." is mentioned. Then, the next sentence, the predictive models seem to be this study's purpose. However, it is unsure what kind of quantity and quality of available data this study can be collected. What "statistical methods" are referring here?
The second paragraph of the Introduction section is hard to understand how it is related to the first and the third paragraphs. Some statements (sentences) have citations but not convincing enough to the need of carrying out this study regarding the need of using the multimodal databases.
The next paragraph is clear and has strengthened the need of using multimodal databases. However, "In the last years" is inaccurate. Please update the writing when the time frame has been shifted.
The rest of the three paragraphs are fine which helps strengthen the argument for using multimodal databases.
Before reading the second section about Methodology, I suggest having a quick and clear review about the prior studies. Then, the method, especially the section 2.3 Data Mining algorithms that your study adopted can be considered better than the ones adopted in the prior studies. In the section 2.2, different kinds of datasets were mentioned, but not sure how they were trained in the model to finally allow authors to report the Results (in the third section).
Tables 1~5 may be placed in the appendices. Section 3.1 may be considered to integrate with section 2.2. For different blocks of databases, e.g., patient personal data, weather information, and sociological information, please provide clear rationales for how the databases were organized and finally be able to create Figure 4. When carefully checking this figure, it is unsure whether the social information is embedded in the "social-economic group" in Figure 4?
Figure 5 seems unnecessary. It still doesn't help me understand how all the databases were linked in Figure 4.
Section 3.2 show the results of the predictive model. However, it is unsure what to predict and for what?
Figure 6 was also created without careful discussion and can only find the conclusion in the last section.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 4 Report
Some more datails about methodology/data mining algorithms migh be valuable - it would e eaiser to understand the procedure ..
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 5 Report
The authors created a database with a great diversity of information, for emergencies that can appear in health activity. The attributes of the database were analyzed and several forecasting models were successfully built.
Remarks
The references are quite old. Mostly references should be within the last 5 years.
Minor
economic, economic - raw 20
Please align the figures
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Thanks for the corrections and other changes. However, if by "Generated Linear Model," you mean "Generalized linear model," it still does not address the concern raised earlier. The sentences describing this seem to have taken from wikipedia without attribution.
Overall, the contribution of the paper is limited and with basic issues like the above, it may not be a good idea to proceed with the publication.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
