A Fuzzy Model for Reasoning and Predicting Student’s Academic Performance
Round 1
Reviewer 1 Report
The manuscript entitled " A Fuzzy Model for Reasoning and Predicting the Student’ Academic Performance" investigated the research works in area of employed the fuzzy concept in student’s academic performance and proposed a fuzzy model for reasoning and predicting the student’s academic performance. It seems to be interesting, but I have some following concerns:
1. Authors should review recent researches related to fuzzy theory with computing with words , such as Nested Probabilistic Linguistic Term Sets.
2. Some parameters need to be further explained, and why choose these values?
3. More discussions should be conducted to show advantages of the proposed method.
4. The layout and the quality of figures/tables of the paper needs to be improved.
As I mentioned before, I think you need to get answers to other kind of questions.
Author Response
Response to Reviewer 1 Comments
1) Authors should review recent researches related to fuzzy theory with computing with words, such as Nested Probabilistic Linguistic Term Sets.
Authors response: We have made major changes to handle these issues (see section 1 pp 2 line 47 and section 2: pp 2 inside the table, pp 6 starting from line 245 to 278)
2) Some parameters need to be further explained, and why choose these values?
Authors response: These points have been handed in different places in the revise version
3) More discussions should be conducted to show advantages of the proposed method.
Authors response: These points have been handed in different places in the section 5, new sub section has been added see page 19 and page 20.
Regarding the Quality of English (Moderate English changes required)
The language has been proofread by professional and all languages issues has solved
Reviewer 2 Report
"Related work" chapter: it would be good to expand the information about the described publications so that they contain not only the name of the method used, but also relevant information, e.g.: number and type of inputs to the method, the result achieved. It would also be advisable to provide a summary of this chapter, e.g. to indicate differences or similarities in the discussed articles.
The algorithms shown in Figures 4 and 5 need to be checked. In particular, it seems advisable to check the loop control variables: "for i =0, i++, i<n; for j =0, i++, j<n;" in Fig. 4 and "for i =0, i++, i<n; for i =0, i++, i<n;" in Fig. 5.
Table 2 - The "Attendance 100%" column header is not consistent with the values in the column, e.g. 0.76
The method of constructing the input fuzzy set requires discussion. The authors used the highest value in the given set for normalization - e.g. in the training set this value is 1, so the fuzzy set is the same as the normal set, the maximum value in the input test set is 0.96 - the fuzzy set is different from the normal set and the maximum value in in this set is also 1. If we forecast the result for the group of lazy students who attended classes only the minimum required number of hours, and for the group of hard-working students, where everyone always attended classes, then for both groups the input (fuzzy) sets will be the same and, consequently, the same predictions should be. Is that a correct assumption?
LINE 306: "The matrix on figure 3" - wrong figure number is given.
The method of determining the value of the FTM matrix requires explanation (do the values in Fig. 7 come exactly from the formula in line 304?). There is no explanation of how to use FTM in forecasting.
What results from the comparison of linear regression and the proposed model? That the model is good, or rather that linear regression will not be able to model the variability of the problem? Perhaps it would be advisable to compare the proposed method with more suitable methods.
Author Response
Response to Reviewer 2 Comments
1- "Related work" chapter: it would be good to expand the information about the described publications so that they contain not only the name of the method used, but also relevant information, e.g.: number and type of inputs to the method, the result achieved. It would also be advisable to provide a summary of this chapter, e.g. to indicate differences or similarities in the discussed articles.
Authors response: We have made major changes to handle these issues (see section 2 starting from page 2 to page 7).
2- The algorithms shown in Figures 4 and 5 need to be checked. In particular, it seems advisable to check the loop control variables: "for i =0, i++, i<n; for j =0, i++, j<n;" in Fig. 4 and "for i =0, i++, i<n; for i =0, i++, i<n;" in Fig. 5.
Authors response: We corrected the error, thank you for observed this point.
3- Table 2 - The "Attendance 100%" column header is not consistent with the values in the column, e.g. 0.76
Authors response: This point has been handed.
4- The method of constructing the input fuzzy set requires discussion. The authors used the highest value in the given set for normalization - e.g. in the training set this value is 1, so the fuzzy set is the same as the normal set, the maximum value in the input test set is 0.96 - the fuzzy set is different from the normal set and the maximum value in in this set is also 1.If we forecast the result for the group of lazy students who attended classes only the minimum required number of hours, and for the group of hard-working students, where everyone always attended classes, then for both groups the input (fuzzy) sets will be the same and, consequently, the same predictions should be. Is that a correct assumption?
Authors response: Is that a correct assumption? No it’s not like that, because:
- Our model did not work one by one value, if it works like that your assumption will be correct, but due to the model approach when building the FTM we multiplied (using AND operation) the A vector (the fuzzy set of the attendance rate) with B vector (the fuzzy set final exam) , then the output will be the first matrix, say Matrix Q , then we multiplied (using and operation) the ¬A vector (the negation of the fuzzy set of the attendance rate) with Y vector (all values are =1) then the output will be the second matrix, say Matrix R, then we add (using OR operation) Matrix Q with Matrix Y to get FTM matrix which represents our production rule (see figure 4 and the formula in page 444). Accordingly, FTM makes the different or we can say it represented the system behavior, not one by one value, i.e this avoids what you think will happen, hence when we need to use the model for predicting new values we can used any data set whatever there are corresponding equivalent to the training data set or not, (therefore we compared our model using linear regression, because FTM play the same rule of the regression line in understanding the system behavior).
- We think this variation between the training and tested data make more challenges in our model and gives us big chance for examine the goodness of our model, because more similarity between the training and tested data will make the system easy, and will provide (by default) strong results (see section 5.2).
- Finally: we sincerely thank you for this comment, (you deeply reviewed our paper), we have added some explanation to declare this issue through the paper.
5- LINE 306: "The matrix on figure 3" - wrong figure number is given.
Authors response: We corrected the error, thank you for observed this point.
6- The method of determining the value of the FTM matrix requires explanation (do the values in Fig. 7 come exactly from the formula in line 304?). There is no explanation of how to use FTM in forecasting.
Authors response: This point has been handed
7- What results from the comparison of linear regression and the proposed model? That the model is good, or rather that linear regression will not be able to model the variability of the problem? Perhaps it would be advisable to compare the proposed method with more suitable methods.
Authors response: These points have been handed in different places in the section 5 and 6, in section 5 new sub section has been added to section 5 see page 19 and page 20.
Reviewer 3 Report
The literature part can be boring if given paragraph by paragraph. Therefore, a fluent summary should be provided. The purpose of this section is to inform the reader about the background of the study. You mentioned who did what. This does not catch the reader's attention and is skipped. In this case, the reader cannot obtain information about the background of the study. Please edit this part.
Secondly, I cannot see the full name of the model in the summary and on the paper. There are a lot of systems called fuzzy systems. I don't understand exactly what is the difference between these and you created them yourself.
Third, how accurate is the data to be given? It would be better if you provide information about the data you presented in the study. It's as if all the accounts were attached to the paper. but the results are only in the table. and only available with MSR. Why is there only MSR? Many parameters such as bias, NSE can be added. Also, how logical is it to give the results only in the table? Taylor diagrams, violin diagrams these can be used...
Finally, I think that the discussion part of the study is missing. you should also indicate what it contributes to the literature and what it provides for future work. I think it's acceptable when these are done.
Author Response
Response to Reviewer 3 Comments
1) The literature part can be boring if given paragraph by paragraph. Therefore, a fluent summary should be provided. The purpose of this section is to inform the reader about the background of the study. You mentioned who did what. This does not catch the reader's attention and is skipped. In this case, the reader cannot obtain information about the background of the study. Please edit this part.
Authors response: We have made major changes to handle these issues (see section 2 starting from page 2 to page 7).
2) Secondly, I cannot see the full name of the model in the summary and on the paper. There are a lot of systems called fuzzy systems. I don't understand exactly what is the difference between these and you created them yourself
Authors response: These points have been handed We have identified our model by using abbreviated name (FPM) as an abbreviation to: Fuzzy Proposition Model, to distinguish it from the other models, we referred to this abbreviated in the summary (the abstract) and we adhered to use it throughout the paper to address this issue.
3) Third, how accurate is the data to be given? It would be better if you provide information about the data you presented in the study. It's as if all the accounts were attached to the paper. but the results are only in the table. and only available with MSR. Why is there only MSR? Many parameters such as bias, NSE can be added. Also, how logical is it to give the results only in the table? Taylor diagrams, violin diagrams these can be used.
Authors response: These points have been handed see the sub section no. 5.3 page number 19.
4) Finally, I think that the discussion part of the study is missing. you should also indicate what it contributes to the literature and what it provides for future work. I think it's acceptable when these are done.
Authors response: These points have been handed in different places in the section 5 and 6, in section 5 new sub section has been added to section 5 see page 19 and page 20,
Regarding the Quality of English (English language and style are fine/minor spell check required)
The language has been proofread by professional and all languages issues has solved
Reviewer 4 Report
Dear Authors,
Please note that the report is attached to this submission.
Best wishes
Comments for author File: 
Comments.pdf
Author Response
Response to Reviewer 4 Comments
- The Abstract should contain answers to the following questions: What problem was studied and why is it important? What methods were used? What are the important results? What conclusions can be drawn from the results? What is the novelty of the work and where does it go beyond previous efforts in the literature? Therefore, please check the abstract for these comments.
Authors response: We have made major changes to handle these issues, see the revised abstract.
- The whole manuscript should be extensively checked for typos and grammatical errors. An overall review is needed for fixing the grammatical and typos errors in the manuscript. For example, the first sentence if the Abstract is not correct. Moreover, please use present simple tense instead of the past simple tense in the Abstract. In page 1, line 40, the word ”development” should be ”developments”. Please check each sentence in your paper.
Authors response: The language has been proofread by professional and all languages issues has solved
- Since you explained the content of Table 1 at the end of Section 2, there is no need to write this table.
Authors response: We think this table provides needed summary and classification to the related work, we think keeping this table on the paper will not affect the paper quality.
- Punctuation marks should be checked throughout the paper, especially after the equations and at the end of the statements.
Authors response: These points have been handed
- Please use a unique style in the list of references.
Authors response: These points have been handed
- In conclusion, I think a minor correction is needed for this paper to be published in the Journal of Applied Sciences.
Authors response: Thank you for your support comment, we hope we did the required correction
Regarding the Quality of English (English language and style are fine/minor spell check required)
The language has been proofread by professional and all languages issues has solved
Round 2
Reviewer 2 Report
Thank you very much for your explanations. I accept authors' answers.
Author Response
Thank you very much for your explanations. I accept authors' answers.
Authors response: thank you for accepting our answers, we appreciate your valued comments that helped us in advancing this research work.
