Next Article in Journal
Kernel Identification of Non-Linear Systems with General Structure
Next Article in Special Issue
Perpetual American Cancellable Standard Options in Models with Last Passage Times
Previous Article in Journal
A Simulation-Based Optimization Method for Warehouse Worker Assignment
Previous Article in Special Issue
Cross-Entropy Method in Application to the SIRC Model
 
 
Article
Peer-Review Record

Feasibility of Kd-Trees in Gaussian Process Regression to Partition Test Points in High Resolution Input Space

Algorithms 2020, 13(12), 327; https://doi.org/10.3390/a13120327
by Ivan De Boi *, Bart Ribbens, Pieter Jorissen and Rudi Penne
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Reviewer 3: Anonymous
Algorithms 2020, 13(12), 327; https://doi.org/10.3390/a13120327
Submission received: 21 November 2020 / Revised: 2 December 2020 / Accepted: 3 December 2020 / Published: 5 December 2020
(This article belongs to the Special Issue Algorithms for Sequential Analysis)

Round 1

Reviewer 1 Report

The comments can be found in the attached PDF file

Comments for author File: Comments.pdf

Author Response

Dear reviewer,

 

Thank you for your time and your critical yet constructive remarks. We believe they helpen to improve the paper.

Please see the attachment.

 

Kind regards,

 

Ivan De Boi

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript “Feasibility of Kd-trees in Gaussian Process Regression to Partition Test Points in High Resolution Input Space” discuss the extension of the Gaussian process regression (GPR) by the concept of kd-threes. The main contribution is the proposing of the several cut-off rules for the exploitation of the kd-tree concept and the combining of them to the new algorithm for application of the concept on GPR for datasets corresponding to the high spatial resolution.

The article provides an introduction in to the concepts of GPR and kd-threes, description of the cut-off rules and of the new algorithm. The manuscript is completed with the presentation of several examples of application of the introduced algorithm on 2D and 3D high resolution datasets and discussion of the results.

The presented work provides the scientific novelty and relevance. The presentation is clear and well-structured and presented experiments show the performance of the introduced algorithm.

Remarks and comments

  • L 110: $L^T\(L\y)$ should be replaced by the corresponding equation.
  • Cut-off rules (9),(10),(11) and application of them: At some point the used threshold values should be indicated
  • Algorithm 1: It would be nice, if at least the most important variables and the object-structure of the node would be mentioned in text.
  • L 234-235: Unity game engine and Math.NET should be cited.
  • L250: Average 2\sigma should be explained.
  • Sec. 4: Were the complete dataset used for the training or was it split into training and evaluation set. If second, then in which relation?
  • References: Different styles are used, par example: year in bold and not bold.

Author Response

Dear reviewer,

 

Thank you for your time and constructive feedback. Please see the attachment for our responses.

 

Kind regards,

 

Ivan De Boi

Author Response File: Author Response.pdf

Reviewer 3 Report

See report.

Comments for author File: Comments.pdf

Author Response

Dear reviewer,

 

Thank you for your time and constructive feedback.

Please see the attachment for our responses.

 

Kind regards,

 

Ivan

Author Response File: Author Response.pdf

Back to TopTop