An Intelligent Hybrid Machine Learning Model for Sustainable Forecasting of Home Energy Demand and Electricity Price
Andrius Tamosiunas
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe research seems interesting from the point of view that the authors apply different techniques and models in combination to obtain better prediction results.
The authors say they have used data from five distinct homes situated across Massachusetts, USA, each equipped with multiple appliances and devices.... And "The dataset comprises energy consumption data for these appliances and devices throughout the year 2016, recorded at various intervals ranging from one minute to one hour.
But, this is unconvincing... let me explain. In my opinion, these are very old data, about 8 years have passed. And only obtained from 5 houses. It is a very very small sample to affirm that the proposed methodology is very valid or better than others.
The authors have two important tasks ahead of them:
First: Carry out the study with a much larger database made up of at least dozens of houses, at least.
Second: Acquire much more recent data, from the last 2 years for example, but not further back.
On the other hand, when we are talking about predicting the demand of such small consumers (households), it is necessary to explain what ways to improve consumption efficiency could be applied at certain times, based on the conclusions obtained. To benefit the users themselves (housing consumers) but also the distribution companies.
The authors must significantly improve the paper.
Comments on the Quality of English LanguageMinor editing of English language required
Author Response
Dear Reviewer,
Thank you for your valuable comments. Enclosed, you'll find our response.
Regards,
Dr. Hamid Khayyam
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe topic is interesting and I enjoyed reading the paper. However, there are some amendments that are necessary before it can be considered for publication.
1. The authors need to state what the contribution of the paper is even in the abstract. Is it the methodology, the variables examined, the problem addressed?
2. Lines 39-40: Worldwide?
3. Lines 63-64: Besides the merit of AI/ ML methods, what is their aim compared to the other two approaches described in the same paragraph?
4. Please separate the literature review from the introduction. The introduction needs to unveil what the exact problem studied is, and what are the novelties introduced. The literature review needs to identify the gap and provide proof that the identified problem has not been addressed so far.
a. Most of the literature focuses on the methodology. What about the relation between demand and prices? Could the number of houses selected influence the total demand and thus the prices?
5. Please define all acronyms before their first use (e.g. ANN is defined at its second use; ARIMA also; CART is not defined). As soon as you define them you do not need to repeat both the acronym and the full name (e.g. ANN).
6. Lines 250-251: “In this study, we utilize one year of data from five distinct home energy demand and prices…” è This sentence seems to be incomplete. Please rewrite it so that it is clear what the dataset is.
a. This paragraph is the most important as it needs to explain what is different from the existing literature. The extensive literature focuses on the methodology but not on the problem. What is truly the contribution of the paper? A new dataset (which is not clear yet what it is)? New variables? New methods (although from the literature presented it seems that one way or another the methods have been already employed). Or is it simply the choice of the optimal method (as stated in this paragraph)?
b. The problem is stated in section 2 that follows, but still there is no indication of what the novelty is.
7. Please explain the rationale behind the choice (and appropriateness) of the selected methods. Why these among the available ML approaches?
8. The authors seem to base their study on 5 houses only. Is this enough? Aren’t the findings heavily dependent on such a small sample? Can they be generalized? What if these houses were in another city, state or country? Is there proof in the literature that such a small number of houses is sufficient?
a. Please also consider this question/ comment in association to comment 4.a. above.
9. Please explain/ introduce the notation used for the houses, before referring to some of them (C, D, F, G).
10. Table 1: What is the measurement unit for each of the variables? What are the predicting variables in each of the houses (only the number is mentioned)? How reliable is the approach when a different number of variables is used for each house?
11. Lines 317-319: Consumers naturally exhibit a preference for utilizing electricity during periods of lower prices rather than when costs surge è If they have the option to do so. In some cases, the use us inelastic (for example in older houses in countries where there are no alternatives).
12. Figure 1: Please define the symbols before their first use (e.g. Pt, Pt-1, P_MA2, etc.).
13. What exactly is mutual information (MI)?
14. Line 449: What is the measurement unit of computational running time (RT)?
15. Lines 457-458: What are the 20 scenarios?
16. Equation (1): What is y?
17. Restrictions/ conditions (2): Could you please elaborate on these conditions? Why are they chosen in such a way? What is their meaning?
18. Please elaborate more on the findings. Assume that a sixth house, called H, enters in the pool – with comparable characteristics and appliances. How can the models presented be readily applied to predict the demand? If more houses enter the pool, will the prices be affected? Is there a critical number of houses (i.e. demand) that will affect the prices?
19. Please mention the source of the tables. It is most likely author estimates, so please do mention so.
20. Please use uniform formatting. For example, the heading of subsection 3.1 is in italics, whereas that of subsection 3.2 is not.
21. There are some grammatical and syntax errors, and typos that the authors are advised to find and correct. In some sentences words seem to be missing.
Comments on the Quality of English LanguageIncluded in the previous comments.
Author Response
Dear Reviewer,
Thank you for your valuable comments. Enclosed, you'll find our response.
Regards,
Dr. Hamid Khayyam
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsI have attached my comments as a PDF below.
Comments for author File: Comments.pdf
Author Response
Dear Reviewer,
Thank you for your valuable comments. Enclosed, you'll find our response.
Regards,
Dr. Hamid Khayyam
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsThe authors have improved and expanded the article considerably. Now the most important points are better justified.
And they have included new references, to clarify the model they have developed compared to the application of other models, previously used by other researchers.
Author Response
Dear Anonymous Reviewer,
Thank you for your valued comments and for accepting our manuscript.
Regards,
Dr. Hamid Khayyam
Reviewer 2 Report
Comments and Suggestions for AuthorsThe paper has been improved and my comments/ recommendations have been successfully addressed by the authors. There are some minor points before the paper can be accepted.
1/ Add the limitation of the choice of smart homes - maybe at the conclusions, as explained in comment C11.
2/ Please check for minor typos, syntax and grammar errors.
3/ There is no separate literature review section (as per comment C4), but if the journal does not impose such a requirement, then the paper can remain as is.
4/ Please mention somewhere that all tables/ figures have been created by the authors as per comment C19. Ideally, under each table/ figure with a simultaneous mention of the data sources.
Comments on the Quality of English LanguagePlease see the comments above.
Author Response
Dear Anonymous Reviewer,
Thank you for your valuable comments.
Here are our responses to your comments:
The paper has been improved and my comments/ recommendations have been successfully addressed by the authors. There are some minor points before the paper can be accepted.
C1. Add the limitation of the choice of smart homes – maybe at the conclusions, as explained in comment C11.
Reply: The method proposed in this paper does, indeed, necessitate a smart home equipped with the capability to measure the energy consumption of various appliances. Modified on page 15, lines 611 – 613.
C2. Please check for minor typos, syntax, and grammar errors.
Reply: The typos, syntax, and grammar are modified throughout the paper.
C3. There is no separate literature review section (as per comment C4), but if the journal does not impose such a requirement, then the paper can remain as is.
Reply: Thank you for your suggestion. We keep the current format.
C4. Please mention somewhere that all tables/ figures have been created by the authors as per comment C19. Ideally, under each table/ figure with a simultaneous mention of the data sources.
Reply: Please see page 14, line 579 – 580.
Reviewer 3 Report
Comments and Suggestions for AuthorsThe draft has been improved significantly based on the previous feedback. The methodology is now clearer, the captions of figures and tables more informative and the obtained results are explained in more depth. Based on this, I recommend to publish the draft in its current form.
Author Response
Dear Anonymous Reviewer,
Thank you for your valued comments and for accepting our manuscript.
Regards,
Dr. Hamid Khayyam
