Carnivorous Plant Algorithm and BP to Predict Optimum Bonding Strength of Heat-Treated Woods
Round 1
Reviewer 1 Report (Previous Reviewer 1)
Dear editor,
Compare to previous analyzed version of the manuscript, I observed important changes, which were not proposed by me. The changes proposed by me were solved. Therefore, in my opinion the final acceptance of the manuscript should be based on the comments of the reviewer that suggested the substantial changes.
Also, I suggest that in the abstract the authors should add in the first phrase that the model was also developed to predict the surface roughness not only for bond strength.
Thank you!
Author Response
Dear Reviewer,
Thank you very much for your willingness to review our paper and for your comments, which helped us significantly improve our paper.
Comment: "I suggest that in the abstract the authors should add in the first phrase that the model was also developed to predict the surface roughness not only for bond strength."
Answer: revised
On behalf of all authors, Ms. Yue Wang
Reviewer 2 Report (Previous Reviewer 2)
The manuscript was significantly improved and can be accepted for publication.
Author Response
Dear Reviewer,
Thank you very much for your willingness to review our paper and for your comments, which helped us significantly improve our paper.
On behalf of all authors, Ms. Yue Wang
This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.
Round 1
Reviewer 1 Report
Dear Authors,
Please find attached my comments regarding your work.
Reviewer
Comments for author File: 
Comments.pdf
Reviewer 2 Report
Carnivorous plant algorithm and BP to predict optimum bonding strength of heat-treated woods review
Dear authors. Thank you very much for the article which is good to read and contains very useful results. The goal of the article was to optimize the back propagation network with the use of the CPA carnivorous plant algorithm.
You have used a very high number of statistical characteristics for comparison of the CPA-BP neural network versus the BP network but there is a very high number of missing information that must be added into the text for eventual publishing of the article which mainly are:
1. Every neural network gives information about training performance, testing performance, and validation performance, and also training error, testing error, and validation error. Why these parameters are not involved in the text or in the Table for comparison of BP and CPA-BP network
2. When given information about R2, MAPE, MSE, and MAE changes using the CPA algorithm you must also give information about typical measurement errors of studied properties. However, if the changes are lower than 3% we can hardly conclude that one network is better than another because the basic variance coefficient for wood bond strength is appr. in intervals between 10% and 20%.
3. If you make predictions also in the Matlab you should also have a prediction matrix of results. These results can be easily compared in Excel for founding if the results of CPA-BP and BP have statistically significant different values of mean and standard deviation. I think if you will make a f-test and t-test or ANOVA all these statistics will give information that means and standard deviation are statistically the same.
Some minor comments:
1. You should give more insight into heat-treated wood, parameters of heat treatment like atmosphere used (air, nitrogen, oxidizing..), temperatures used, time, etc., and how these parameters influence the physical and mechanical properties of heat-treated wood. There are plenty of recent publications in this area like (10.1007/s10853-020-05722-z, 10.1080/17480272.2021.2014566, 10.1080/17480272.2018.1450783, 10.3390/polym12020485, 10.1515/hf-2017-0205, 10.3390/f12020249).
2. It is possible to conclude that in all studied wood materials is CPA-BP
ANN network better as BP network?
3. Which are the variance coefficients of bond strength measurements for all
studied parameters? Can they be represented by one unique variance error?
4. At which confidence interval are results compared? It is a 99% confidence
interval or 95% confidence interval
5. All statistical parameters should be divided for every material type.
