A Data Factor Study for Machine Learning on Heterogenous Edge Computing

Round 1

Reviewer 1 Report

The authors describe in the article an agricultural crop monitoring environment constructed based on sensors and using Raspberry Pi and Arduino as the edge computing layer. The collected data is processed in the edge computing layer using machine learning techniques. Unfortunately, the article does not have a clearly defined scientific purpose of the research. What is innovative in the proposed solution? The use of sensors and edge computing devices to process data using machine learning algorithms is nothing new in the Internet of Things. The authors did not even mention their original achievements anywhere in the article. The proposed environment was also not compared with other such environments, nor was it shown how it would be better than other devices of this type. The drawings are of very poor quality - they look like poor quality scans from paper versions. In addition, the English language is at a very low level. It is impossible to understand the meaning of most sentences. The results of the research and the quality of the paper are not sufficient for a journal.

 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

Reviewer’s Report on the manuscript entitled:

A Data Factor Study Using RMSE for Machine Learning on Heterogenous Edge Computing

 

The authors perform an experiment that automatically collects relevant data via detection devices and performs data modeling and calculation in an edge computing environment. Though the topic and results seem interesting, but the manuscript is very poorly structured and written and requires extensive revisions. There are many grammar issues in the manuscript. Below please see my comments.

 

Please avoid using acronyms in the title. Please write “Root Mean Square Error” instead of RMSE.

 

Line 13. “For intelligent agriculture, use the IoT to collect data.” This is a rather strange sentence and does not align with the flow of Abstract. Please use a longer sentence and more informative. Also, please define IoT (Internet of Things). Please define all the abbreviations the first time they appear.

 

Lines 49-51. Grammar issue. What is the subject of the sentence? Please rewrite.

Line 60. Please remove “From”.

Line 65. “it can”? What can?

Line 68. Here you should highlight the main contributions of the manuscript using bullet points.

Line 96. “To improve optimisation of deep learning models”. Grammar issue. This is a very poor way of writing. The manuscript must undergo extensive English grammar revisions.

 

Line 180. Deep Learning Restricted Boltzmann Machine has been utilized for energy consumption forecasting, while considering factors, such as temperature, humidity, etc. which can also be included here:  https://doi.org/10.3390/su141610081

Also, in https://doi.org/10.3390/su141811163, deep learning-based neural networks are applied to understand both the linear and non-linear relationships between the desired variables and identify the causal effects.

 

The manuscript must be restructured.

Sections 2.1, 2.2 do not belong here as they are related work and literature review. They should move to Section 1. Please note that Materials and Methods should describe your datasets and methods/formulas only. Section 3.1 belongs to Materials and Methods as it describes the datasets. Sections 3.2 and 3.3 seem to be your Methods. Then what is your result?

 

 

Figure 6. Please avoid screenshots that refer to a specific computer code.

 

Figures 4, 7, Table 1, 2 should be in table format (typed) not screenshot. Figures 4 and 7 must be called as Tables.

 

Figures 10, 13. In the bottom left corner, please remove these numbers. Why should MSE have so many decimals? This figure is poorly and carelessly generated.

 

Line 335. “his”!?

 

Line 365. Please rewrite. What was the objective and what did you propose. Please use past verbs in Conclusions. For example, “In this paper, we proposed…”.

Lines 371-372. Grammar issue. Please rewrite.

Line 375. Just use “IoT” here instead of “Internet of Things”.

 

Please carefully check the references to ensure they are correct and have a consistent format according to the MDPI guidelines.

https://www.mdpi.com/journal/applsci/instructions#preparation

 

 

There are many other typos/punctuation/grammar issues that I did not mention. They must be corrected.

Regards,

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

The paper experiment automatically collects relevant data via detection devices and performs data modeling and calculation in an edge computing environment. At the same time, the data model is transmitted through the communication protocol, providing another node to verify the modelling and calculation results. Experimental findings demonstrate that the single-point data formation model can accurately pre-21 dict the growth trend of plants. The abstract needs to be more attractive by adding some implications and applications of the results.

 

In experiments, one retrains and learns on the same data set. During the first learning, the average square error for this value was 7.8. However, this is not very clear from the experiments. Although the results of formation and prediction of the same parental data are very accurate, the average difference is too large. Although this can be attributed to plant growth characteristics, environmental data may not directly correlate with growth trend results. Please give more explantions.

 

It is always possible to effectively predict that outcomes are consistent with initial outcomes if the same dataset is used. In the study, the dataset training outcome model was applied to the second non-homologous training test object, and the predicted calculation results and actual data are presented in Figure 10. There are no evaluations provided. 

 

Research is rebuilding IoT data collection and training model under advanced computing. Based on every edge note there is a processor kernel in the raspberry pie environment. In our experiments, we use sensor data collection and simulation training to develop relevant data and results for analyzing foliar growth trends. What are the potential limitations of this? 

The writing of the paper is poor. I recommend using a professional editing service if possible. 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Unfortunately, the authors have not addressed any of the crucial issues. The article does not have a clearly defined scientific purpose of the research. What is innovative in the proposed solution? The use of sensors and edge computing devices to process data using machine learning algorithms is nothing new in the Internet of Things. The authors did not even mention their original achievements anywhere in the article - probably because there are no such original contributions. The proposed environment was also not compared with other such environments, nor was it shown how it would be better than other devices of this type. The drawings are of very poor quality - they look like poor quality scans from paper versions. In addition, the English language is still at a very low level - there are sentences hard to understand. The results of the research and the quality of the paper are not sufficient for a journal.

 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

Thank you for improving the manuscript. Some of my comments were not addressed, however. Please see below my remaining comments:

Please make a brief discussion of my point 9 with references I suggested. Line 444-451. Please move this to the end of the discussion section and include the references I suggested before which discuss these in detail.

Please modify the title as: "A Data Factor Study for Machine Learning on Heterogeneous Edge Computing"

Line 70. Please describe how the rest of the paper is organized.

The paper is poorly structured. Sections 2.1 and 2.2 should not be in the Materials and Methods. They can be called "Related work" and "Research objectives" or "Research Contributions"

Line 356. "Figures 5,6, and 7" not Figures 5, Figures 6, Figures 7...(Grammar issue). Line 357. "Figure 6". Please check these issues elsewhere.

The format and style of references must be according to the MDPI guideline.

Please carefully proofread the manuscript.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

The paper has been revised. I recommend publication. 

Author Response

We thank the reviewer for the positive comments.