Semantic Recognition of Ship Motion Patterns Entering and Leaving Port Based on Topic Model

Round 1

Reviewer 1 Report

This article proposes a method for a semantic recognition based on a topic model for ship motion patterns, and more particularly for the patterns entering and leaving the ports. This method has been applied to the Chinese port of Tianjin. Using an unsupervised method, it considers the influence of port rules such as navigation channels and dedicated areas, in order to discover patterns that are representative and interpretable.

The article is interesting, well curated and written, and proposes an interesting and innovative method, including an application of AIS inside ports, which is not that common.

I recommend the publication of this article, provided that some minor revisions be done.

Some comments:

You chose to use ArcGIS, which is a proprietary software, and ENCs, which are not publicly accessible. Ideally, this has to be avoided. Can you justify your choice? Why not use open source alternatives such as OpenSeaMap?

Line 89, you refer to the datAcron project. I rather suggest one of those two papers be cited, as they are more relevant to the citation of this specific research project than [25], as they are from 2019 and not 2017
10.1109/OCEANSE.2019.8867518
10.1109/OCEANSE.2019.8867554

Lines 176 and 181, you use 0.5 and 3 knots as threshold values. Do you have any literature to support the choice of such values? It seems rather arbitrary.

Table 1 is presented in such a way that the date is a native feature of AIS. This is not true, AIS messages are timestamped by the parser upon reception, but there is no native timestamp within the AIS message broadcast.

Also, in Table 1, MMSI are not anonymised. It could be a possibility to protect privacy.

Table 2 is extracted from AIS technical specifications. A reference should be provided here.

In Table 3, open brackets are figured as (0.5,3], however it should be ]0.5,3]

Line 245, modify “gri1kmX1km”

Please check the equation at lines 350-351

Please precise the base of the log at line 391

Lines 422, 433, 434, numbers are figured incorrectly. 1956 should be written 1,956 and 45438: 45,438. Please check all across the document

Table 5, MMSI and vessel names are not anonymised. It could be a possibility to protect privacy.

Lines 474, 475, 477, and Table 6, MMSI are not anonymised. It could be a possibility to protect privacy.

Author Response

Dear Reviewer:

We are most grateful for your detailed review and most constructive comments and suggestions. We have invested a significant amount of effort again to implement these comments and suggestions. The paper has now been revised to consider all these comments. A revised paper and our detailed response to the reviewers' comments are attached to this letter.

Author Response File: Author Response.docx

Reviewer 2 Report

Title : Semantic Recognition of Ship Motion Patterns Entering and Leaving Port based on Topic Model

Authors : Gaocai Li, Mingzheng Liu, Xinyu Zhang, Chengbo Wang, Kee-hung Lai and Weihuachao Qian

-----------------------

In this paper, the authors present a semantic recognition method based on a probabilistic topic model for ship entering and leaving motion patterns. The method includes three modules: trajectory preprocess, semantic process, and knowledge discovery. 

 

Strengths:

. The abstract presents correctly and synthetically the paper. 

. The ideas proposed in this paper are very interesting and merit to be published. 

. The bibliography part has been worked, and is well analysed in the introduction section. 

 

Weaknesses:

. The 17 topics are not precisely defined, and this hurts the correct understanding of the section 3 examples, yet correctly illustrated. This could be easily corrected. 

. The semantic recognition method should be presented and the statistics methods behind would be very interesting to read. 

. Conclusion could gain in being extended. As well, future works and ideas could be more developed. 

 

References :

. 39 research references, out of which 3 self-references, giving an acceptable self-reference ratio equal to 7.7%.

. The bibliography of this paper is mainly composed of recent references: 7 of them are more than 10 years old, when 32 of them are less than 10 years old.

. Avoid citing groups of references: [2-4], [28-30], but on the whole, references are correctly analysed. 

. The references are cited in the text adequately and appropriately.

 

Originality / Novelty / Scientific soundness:

. The question set in this paper is original and well defined. 

. The study has been correctly designed, and is technically sound. 

. The subject addressed in this paper is relevant, and very useful in busy port areas. 

. The semantic recognition method is insufficiently described. As well, data used are unavailable for readers. Consequently, another researcher would be unable to reproduce the results or to propose anyelse method with the description given in this paper. 

 

Overall evaluation:

. Nevertheless, I think there is an overall benefit to publish this work. 

. In my opinion, methods and conclusions of this paper seem to be interesting for the readership of the journal. 

. As well, this work provides a clever method to address a current problem, clearly highlighted in the abstract. 

. The English language quality and style of this paper are correct, appropriate and understandable. The misprints are understandable, and will be easily addressed.

As a conclusion, my suggestion to the editor is to propose to the authors to perform on their paper the minor revisions suggested in this review before a possible publication in JMSE.

 

--------------------------------

Typos / Comments / Remarks:

--------------------------------

. Lines 153-156: Two times the same sentence… Please correct it into one, complete sentence.

. Figure 1: Only normal-speed sailing ships are considered in figure 1. What about the two other classes (Maneuvering and Hoteling)?

. Line 343: document m --> m in italic.

. Lines 473-474: Sentence in those lines is statistically incorrect, because the sum of percentages on a line is 100. Thus, it can not be 82.1% of the remaining… The sentence formulation is misleading.

. Figure 11: This figure is very interesting. Anyway, it should be precised. Indeed, patterns 7, 10 and 16 all stand for "Container ship leaving the port". It is uneasy to understand the difference.

. Table 6: As well, precise words used to defined the 17 topics of Table 6 should be clearly precised in the text. --> Moreover, the 17 topics are in the caption of figure 12.

Author Response

Dear Reviewer:

We are most grateful for your detailed review and most constructive comments and suggestions. We have invested a significant amount of effort again to implement these comments and suggestions. The paper has now been revised to consider all these comments. A revised paper, test result, and our detailed response to the reviewers' comments are attached to this letter.

Author Response File: Author Response.docx