Regional Logistics Demand Prediction: A Long Short-Term Memory Network Method

Round 1

Reviewer 1 Report

The paper deals with logistics demand prediction by applying neural network modelling techniques.

An interesting work approached under novel methodological modeling techniques. However, there is a confusion in the clear definition of the main objective and aim of the research carried out. Although the authors write which are the main contributions of their work (as for example state in page 2 from lines 68 to 87), in the abstract and in the section of conclusions, they only focus on which is the appropriate model to predict the demand.

The section of the methodology followed is systematic although sections 3.3 -3.5 seem like a literature review. There is an extensive and detailed analysis of the modeling procedure supported by modeling results. The section of the conclusions needs improvement (see comment above).

Below please find some minor points that should be corrected:

The paper requires some editing in respect to the English language.

Line 69, please add also the country.

Line 70, “novel aspect for selectiing indicators…” please correct to “novel aspect for selecting indicators…”.

Line 153, “introduced by John hopfied, please correct to “introduced by John Hopfied…”.

Lines 233, 237 and 243. There are some missing texts, replaced by some text boxes. I think it could be due to errors when writing the equations.

In Tables 1 and 2, I suggest the authors to replace x1, x2, …, x5 as the labels of the first row with the names of the factors/variables (e.g., GDP, income, etc.).

Considering the above, the recommendation for publishing the paper must be reconsidered while the main objective(s) of the research are not clearly stated and give a sense of vagueness for the reader.

The paper could be published as long as the above stated comments are followed.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

This study uses the Long short-term memory (LSTM) network for predicting the regional logistic needs. The Changsha logistics needs prediction index system is introduced and compared with other methods (regression, neural networks, etc.). I propose the following changes/additions to the article:

 

1. Only the data for the 2018-2021 period has been considered in Figures 5, 6, and Tables 3 and 4. Why the entire data has not been considered for such analysis?

2. Elaborate more on the managerial implications in the conclusion section. While discussing these implications, provide an analogy with other disruptive events that may potentially trigger the use of the proposed models/findings.  

3. The literature review section can be extended by discussing the predicting models for logistics demand delivery (deep network, fusion network, etc.), and the inherent flaws in using neural networks. 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

See attached. Minor/Moderate Revision recommended.

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 4 Report

Minor observations

Line 70. Correct “selectiing”.

Line 153. Correct “hopfied”.

Line 184. Replace “to” with “and”.

Lines 186-187. Define: f, sigma and b.

Lines 221, 222. Correct to be inline.

Line 227. Correct “state.. As”.

Lines 236, 243. What is the rectangle? Replace it with the appropriate sign.

Line 322. “the prediction” is with “The”.

Line 327. Correct the legend: “valur” and size of letters.

Figure 5 is not announced before.

Lines 371 to 381. Their place is somewhere before MAE, RMSE and MAPE are first time mentioned, it means before Table 3.

Line 393. Correct size of letters.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

No further comments.