A Data-Driven Forecasting Strategy to Predict Continuous Hourly Energy Demand in Smart Buildings

Round 1

Reviewer 1 Report

The manuscript presents data-driven forecasting techniques to predict hourly energy consumption in buildings. My comments are as follows:

1) One major remark is that, the authors do not present a literature review on building energy consumption forecasting. It is important to present it and then clearly highlight how their paper complements the literature.

2) Overall writing of the paper is average, but in some parts, a thorough rephrasing of the text is needed.

3) Figures quality is not up to the standard of a journal paper.

4) In addition to presenting the results, it is important to explain that why such results are obtained. For example, it could be explained why MLR and ANN methods exhibit poor performance.

5) The authors should show, not only the prediction error figures but also the actual forecast.

6) Why do authors present many different accuracy metrics for comparison in the paper, if they are only using MAPE to compare?

The rest of the comments are in the PDF file. 

 

Comments for author File: Comments.pdf

Author Response

The answer is in the attached document

Author Response File: Author Response.pdf

Reviewer 2 Report

This paper proposes an energy consumption forecasting strategy. The topic is important. However, the following problems must be handled.

Major comments：

1. Please make more discussions that why choose the climatic variables listed in the paper. Besides, the explanation of variable RH2M in Figure 2 is missing.
2. Please explain Figures 2, 3, and 5 in detail.
3. Please make more discussions that why choose the eight forecasting models in Section 3.1.
4. Clarify the main achievements of the paper in the Conclusion section.
5. Section 3.1: the author said “ Based on previous research [14] machine learning and deep learning models were used”. More recent references must be cited to show the rationality of the used models. Plz refer to (Renewable Energy, 2020, 154: 598-613; https://doi.org/10.1016/j.energy.2021.121756; https://doi.org/10.1007/s42452-020-2024-9).
6. This paper only uses the data of two buildings, which is not convincing. So, if possible, please add one or two data sets to further verify the proposed strategy.

 

Minor comments：

1. The expression “…for any hour of the day for the next 24 hours…” is very confusing, does it mean “…for any hour of the next 24 hours…”? Please explain or modify it.
2. It would be better to integrate the hyperparameters of the models in Section 3.2 into one table.
3. The writing should be improved. For example, the sentence "For this research was used the data of two buildings located on the campus of the 84 University of Valladolid, Spain." in Section 2.1 is a sick sentence.
4. Plz check if Formulas 3 and 4 are right? You can refer to the following study. [r1] Forecasting energy consumption and wind power generation using deep echo state network [J]. Renewable Energy, 2020, 154: 598-613.

Author Response

The answer is in the attached document

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Thanks for revising and resubmitting the paper. The paper has been much improved from the previous version. The content of the paper is good, however, I would still suggest the authors to have a thorough review of the paper. There are numerous editorial errors (grammatical mistakes and non-technical phrasing). These errors reduce the quality of the work and make it sound a bit unprofessional. Other than that, I am satisfied with the paper. 

Please find some minor comments (related to editorial errors) in the attached file.  

 

 

Comments for author File: Comments.pdf

Author Response

Thanks, all is corrected

Author Response File: Author Response.pdf

Reviewer 2 Report

The revision is satisfactory.

Author Response

Thanks