Enhancing Driving Safety Evaluation Through Correlation Analysis of Driver Behavior

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The title of the article needs to be rethought. I suggest that removing “Electric Vehicles” from the title since there is no clear evidence that the electric vehicle usage contributes to the behavior change. Also, this study only used one car from once auto maker, and hence, this should not be generalized to all electric vehicles.

In the abstract, FB-growth should be spelled out and when first mentioned in the text. Also, the frequent pattern growth algorithm should be referenced properly when mentioned in the text.

Also, in the abstract, I do not see a benefit of mentioning that SPSS is used. I see this as a weak point as will be detailed below.

The argument in Lines 33-35 needs to be clearly supported. I think what the author is trying to say is that there is a difference between driver behavior in gas-powered vehicles when compared to electric vehicles. Otherwise, the language should be changed to focus on the driver safety evaluation.

The following papers are missing when discussing the driver interactions with the surrounding environment in Line 42.

• A context identification layer to the reasoning subsystem of context-aware driver assistance systems based on proximity to intersections
• Location-based analysis of car-following behavior during braking using naturalistic driving data

The paragraph starting at Line 53 is so long and it talks about three different ideas. This paragraph should be split into three paragraph with each paragraph discussing only one ides to enhance reader experience.

I disagree with the statement made on lines 102 and 103 that talks about traditional cars can not collect driver behavior data. There are hundreds of researches and naturalistic data collection efforts that were done in this area. The papers mentioned above are just examples for traditional vehicle data collection. The following is an example of online data collection:

Connected vehicle V2I communication application to enhance driver awareness at signalized intersections

The authors should discuss how were the drivers prepared to the road trip. Were the drivers aware of the study and its objectives?

The authors used three different software packages (Python, MATLAB, and SPSS). Why? All the analysis can be done on Python alone or MATLAB alone.

Figure 7 is better to be in a bar chart or histogram format with the units of the speed displayed on the figure to avoid any confusion.

It was mentioned that the data was collected during the off peak and there is no mentioned about this again until in the limitations to be included in the future research. However, it would be appropriate to shed light on it here.

All the captions of the figures should be reviewed to be more informative and convey a meaning to the reader about the figure.

Also, the quality of the figures should be improved. For example, Figure 11 can not be read.

Line 387 is very briefly discussing outliers. The followed approach should be referenced. What are the values of the rotated loadings if additional PCs are included.

More explanation is needed for Fig 14. Why there are 6 dots in the plot while the there are 12 rules in the table? The caption of this figure is not informative at all.

For section 4.1.3, there should be a comment on the female sample size and its impact on the results given that the majority of the drivers were male. This point impacts many of the conclusions and finding such as accelerator pedal usage, etc.

The comprehensive weighting in Line 485 should be expanded with an example of how the scores are calculated.

Is fig 21 for the whole sample or for the selected 6 drivers only?

Line 532, the results are aggregate, so this sentence needs to be revised to reflect that. i.e., the results can identify potential unsafe drivers or unsafe trips but will not identify unsafe event within a trip. This is how I understood the analysis.

In conclusions, line 542 is not 100% accurate and needs to be revised and be less generic. Two or more unsafe behaviors "in terms of the selected measures" since the measures are highly aggregated as in my previous comment.

How the proposed approach can be used for fleet management? My understanding is that all the drivers are combined in the analysis.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The topic of the article is interesting. The list of cited articles is relevant, 18 out of 41 cited articles published in the last 5 years. However, I have some comments regarding this article. The authors should describe more clearly what the idea of the FP-growth algorithm is and how this algorithm differs from other algorithms. The novelty of the proposed solution method should be described in more detail. The authors should clearly describe what is novel about their approach to solving the problem and provide a comparison with the results obtained using other methods. In some places in the article, there is a verbal description of formulas (for example, formula 7; lines 380-389). This should be corrected and the mathematical formulas should be included in the article (for example, lines 380-389). For mathematical formulas, it is necessary to describe in detail the area of their application and comment on all input parameters  (for example, formula 8, F1,F2). Figures 12 and 13 have poor quality graphics, and they should be redone and detailed comments should be given. The article can be published after significant corrections.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Editor,

 

I wish to thank you for the opportunity given me to review this manuscript entitled “Enhancing Driving Safety Evaluation through Correlation Analysis of Driver Behavior in Electric Vehicles”. Please find below my review comments:

1. The abstract section should include the main aim of the study.
2. The introduction section is too long, please reduce.
3. I would like to suggest that the authors should move the experimental design, data collection and data processing sections to the “Methods and Materials” section and reduce the descriptions.
4. Some of the Figures such as Fig. 8, 11 are not clear. Please enhance the quality of the Figures.
5. Generally, the paper is too long. Reduce the texts and the descriptions in the “Methods and Materials” and “Results and Discussion” sections.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Figure 14 is showing 12 points while the response says there are 14 points. 

Author Response

Comment 1: Figure 14 is showing 12 points while the response says there are 14 points.

Response 1: Thank you very much for your careful cross-checking. We sincerely apologize for the inconsistency in our previous response. In our earlier reply to Reviewer 1’s Comment 15, we stated: “We would like to clarify that Figure 14 actually visualizes 14 association rules, with each dot representing a rule and colored by its lift value. In the initial submission, PDF compression and the original colormap resulted in some dots appearing faint, which may have caused confusion.”

This was an incorrect explanation on our part. In fact, Figure 14 displays 12 association rules, not 14 as previously stated. We apologize for the mistake and have corrected it accordingly.

More importantly, we confirm that the revised manuscript content was accurate, as it consistently described the number of rules analyzed in this figure as 12. To avoid further confusion, we have highlighted the correct statement in the revised manuscript (lines 444-446) in yellow for easy reference by readers and reviewers.

We truly appreciate your detailed and thoughtful review, which helped us improve the consistency and clarity of our submission.