Modelling the Publishing Process of Big Location Data Using Deep Learning Prediction Methods

Round 1

Reviewer 1 Report

This paper proposes a method for modelling the publishing process of location data aggregates using machine learning. The method is innovative while relying on proven techniques. The analysis of the method and its analogues in literature is carried out experimentally, which is a good way to provide an initial appraisal of the method's value.

The paper would be even stronger if a validity analysis were conducted, including external validity and construct validity, to further support the conclusions drawn from the experimental analysis. Overall, the paper presents an original method and a sound analysis that, in addition to being of interest to  location data aggregators, is going to be useful as an application of mixed machine learning techniques.

Author Response

Dear Reviewer,

We express our deepest gratitude to your good self's highly valuable comments on our paper which have been taken into account in the enclosed revision. Kindly note that every attempt has been made to address the comments raised by your good self.

We have accordingly explained point-by-point on how the issues have been addressed in the revised version including details of the changes that were made in the paper. We hope that these changes would satisfy the requirements subsequently making the article acceptable for publishing in the MDPI’s Electronics.

Best Regards,

Yan Yan (Corresponding Author)

Author Response File: Author Response.pdf

Reviewer 2 Report

First all... Good work!

Things to improve:

1) line 63. Suggest split 'Randomness and Uncertainty'  to two different items

2) Definition3 has too many hyper parameters. Needs a bit discussion on how to choose alpha, beta, theta and etc.

3) In literature review, previous work are reviewed but the research GAP is NOT highlighted.

4) Some of the stuff in section 4 and 5 can be moved to lit review.

5) It was not clear that CNN and CNN-LSTM, shown in Table 2 was part of your evaluation, rather than part of your design, until later paragraphs in your article. Please re-organize this article if needed. 

6) Needs a bit more explanation why MODWT over LSTM+GRU is better

 

Author Response

Dear Reviewer,

We express our deepest gratitude to your good self's highly valuable comments on our paper which have been taken into account in the enclosed revision. Kindly note that every attempt has been made to address the comments raised by your good self.

We have accordingly explained point-by-point on how the issues have been addressed in the revised version including details of the changes that were made in the paper. We hope that these changes would satisfy the requirements subsequently making the article acceptable for publishing in the MDPI’s Electronics.

Best Regards,

Yan Yan (Corresponding Author)

Author Response File: Author Response.pdf