You are currently viewing a new version of our website. To view the old version click .
MDPIMDPI
Search all articles

Review Reports

Appl. Sci.2026, 16(1), 296;https://doi.org/10.3390/app16010296 
(registering DOI)
by
  • Lin Wu1,
  • Ahmad Yahya Dawod1,* and
  • Fang Miao2

Reviewer 1: Anonymous Reviewer 2: Anonymous Reviewer 3: Anonymous

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Respected authors,

The submission is of high quality so I must congratulate you all.

I have some sugestions,

Fig 2, I believe it is kernel and not kernal, also check aspect ration (horz vs vert)

Some formulas contain a square symbol that might be caused by a font not included in the pdf. Those apperar in some equations and also in the text.

Please include more explanation about the sampling period, it is the maximum allowed by the hardware?

The main contribution of the paper is the improvement in predictions by means of neural processes combined with optimization. The topic is well presented and within the scope of the publication.

It must be praised the fact that task is not treated as whole black process, i.e. supervised learning of a huge network. Instead the feature extraction is treated with the CNN in the realm of distributed data.

Readers might appreciate this sentence to be elaborated a bit more and with some bibliographic suppor (page 7, line 292) "...excel in handling structured data with local correlations, and their translational invariance,"

The conclusions could be enhanced pointing out future research lines suggested by the actual results of the present investigation.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors
  1. This article appears more like a course lab report than a formal academic paper.
  2. The "Related Work" section contains excessive technical details; only citations should be retained for brevity.
  3. The comparative analysis would benefit from including traditional non-deep-learning methods and directly relevant prior works to strengthen the persuasiveness.
  4. Some experimental results inadequately support the conclusions. For example, the conclusion in Section 4.2.4 conflicts with the results shown in Figure 10.
Comments on the Quality of English Language
  1. The manuscript requires comprehensive revisions to comply with the journal’s formatting guidelines.
  2. Mixed use of full-width/half-width characters persists, and some sentences lack proper punctuation (e.g., P24-4.2.3).
  3. Unexplained underscores (e.g., _CNN) precede certain abbreviations; the purpose of this notation must be clarified.
  4. Figure captions and annotations exhibit inconsistent formatting, with alignment issues observed.
  5. Mathematical equations display disorganized formatting, negatively impacting readability (e.g., P13).
  6. Inaccurate terminology is used for domain-specific concepts.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Authors need to define the research problem and scientific novelty of the study, not just a combination of models.

In the literature review, avoid general references without reasoning, for example, [16]–[27].

Methodology section.

There is no explanation why BKA was chosen among other metaheuristics (e.g., PSO, WOA, GA).

There is no clear objective function of the model (the optimization criterion is only mentioned as MSE).

Some formulas are given without units or parameter designations.

The authors described only some of the hyperparameters (CNN + BKA ranges), but there is no full description of the model. BiLSTM hyperparameters, as well as training parameters, are not given, which makes it difficult to reproduce the work.

Results section.

There is no description of the number of records, time discretization, filtering, and normalization.

It is not specified exactly which parameter is predicted (temperature or humidity).

The sample size is not described: how many time points, what period, what frequency of collection.

There is no description of data normalization, noise processing, omissions.

The figures are descriptive, without quantitative analysis.

The discussion section is practically absent. The authors repeat the results, but do not analyze them causally; there is no critical analysis of the limitations of the model; there is no comparison with the results of similar studies by other authors.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

  1. The main research thread of the paper needs to be better organized, and the overall structure requires adjustment. The connections between chapters should be strengthened. For example, Section 2.4 in Related Work includes the proposed model’s structure and performance description, which would be more appropriate in the methodology/experiments section. Section 3.1 in Methodology explains the reasons for choosing BKA in excessive detail (Table 2, Table 3, Figure 1, Figure 2, etc.). It is recommended to condense this section to highlight the core motivation and advantages. Other parts of Chapter 3 have similar redundancy; overall compression and reinforcement of the key content are suggested.

  2. The proposed method needs a more intuitive introduction. It is recommended that Chapter 3 begin with an overview of the overall model framework/process (including a framework diagram), followed by explanations of each component. This would improve clarity and comprehensibility.

  3. Sections 4.4 and 5 contain a substantial amount of duplicated content regarding model mechanism interpretation, limitations analysis, and future work discussion. This repetition weakens the structural compactness and logical progression. It is recommended to clearly define the purpose of each section, merge or separate redundant content, and strengthen the distinct focus of each part.

  4. The references need to be enriched, particularly with works from the last three years, in order to reflect the latest research progress in the field. It is recommended that the authors review recent frontier studies and supplement the reference list accordingly.

  5. The quality of figures needs improvement. Some image fonts are too small and the resolution is insufficient (e.g., Figures 16, 18, and 19), which affects readability.

  6. The layout of the paper requires further refinement. For example, in line 119, a punctuation mark appears alone on a separate line. Large blank areas appear on pages 23 and 45. Table 1 appears on page 2, but it is first cited on page 26.

  7. The writing needs further standardization. For example, in Table 1, the meaning of “Gao” in the “resource efficiency” column is unclear, and letter capitalization is inconsistent across the table. Line 482 appears alone on a separate line. In line 714, the equation numbering format is inconsistent with the preceding sections. In line 727, confirm whether the notation “Xi” is correct. The caption format of the figure in line 1336 is inconsistent with other figures. The description of the confusion matrix in lines 1276–1288 on page 37 needs to be expressed more rigorously and precisely.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

I recommended publishing the article in present form

Author Response

Thank you for your positive feedback and recommendation. We are pleased to know that you find the article acceptable for publication in its present form and will proceed accordingly.

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

All the major issues have been thoroughly revised.