Fuzzy MLKNN in Credit User Portrait^†

Round 1

Reviewer 1 Report

Interesting study, perhaps the authors should improve the labels from some pictures for better clarity for example the x axis labels in figure 6.

Author Response

Thank you very much for your support.

We have changed. And the changes can be found in Figure.4, Figure.5 and Figure.6. 

Reviewer 2 Report

Good work.

Author Response

Thank you for your affirmation.

Reviewer 3 Report

The paper proposes an improved fuzzy MLKNN multi-label learning algorithm based on MLKNN. The authors provide intuitionistic fuzzy numbers and improve the algorithm by using the corresponding fuzzy Euclidean distance to realize the multi-label portrait of credit users.

 Although the paper is well written and introduces an improved methodology, I have the following concern with the paper.

1- The literature review section should add a table to provide insight regarding the existing research limitations and strengths.

2- The paper should provide a flow chart to explain the selection and ignore criteria of features for readers

3 - The paper should explain why marital status is Strikethrough in Table 3.

4- The authors should compare the results of the proposed solution with the existing solution on various benchmarks identified in the existing literature.

Author Response

1- The literature review section should add a table to provide insight regarding the existing research limitations and strengths.

Response：Thank you very much for your suggestion. The analysis of advantages and disadvantages of existing research algorithms has been added into the literature review section, which is marked in blue.

2- The paper should provide a flow chart to explain the selection and ignore criteria of features for readers

Response：Thank you very much for your suggestion. Feature selection criteria have been added to the relevant parts of the original paper. The specific contents are marked in blue, some of the texts are "First, after removing privacy variables, such as ID number, telephone number, address, etc., the correlation test of variables was carried out, and some variables with correlation exceeding 0.7 were removed......”, because the process is simple, meanwhile, it is not the main contribution of the work, so we think there is no need to add a flow chart.

3 - The paper should explain why marital status is Strikethrough in Table 3.

Response：Thank you very much for your suggestion. The Strikethrough is only in our initial trials, which is off the modified version.

4-The authors should compare the results of the proposed solution with the existing solution on various benchmarks identified in the existing literature.

Response：Thank you very much for your suggestion. The comparative analysis has been supplemented in Section 4.3.

Reviewer 4 Report

Reviewer comments for

Paper Title: Multi-label learning with credit data based on Fuzzy MLKNN

 1) This work shows an updated Fuzzy MLKNN multi-label learning method based on MLKNN to address the subjective enhancement induced by credit data discretization and the lack of a multi-dimensional image of credit users in current credit data research.

Intuitionistic fuzzy numbers reduce credit data subjectivity after discretization. However, employing fuzzy Euclidean distance to create a multi-label credit user portrait improves the method. Fuzzy MLKNN outperforms MLKNN on credit data, with the most improvement on One Error.

3) Block diagram of the complete system can be added to get more flow clarity of the proposed system.

4) discretization and the fuzzification of the data can be elaborated more in section 4.2 with theory behind it.

5) Equations provided suitably. Kindly check if all variables are named.

6) References of recent years published papers seen added.

7) in This paper considers few indicators at the same time. In Figure 3, blue line is the RankingLoss with MLKNN, and in Figure 4 green columnar presents OneError values with MLKNN. And it is shown that both are better for proposed system. Also coverage is used for comparison.

Conclusion section has mentioned this.

8) Please check the journal format.

9) Check for grammar and punctuation etc.

10) Future scope is also provided as it needs more theoretical background or study.

Comments for author File: Comments.pdf

Author Response

For (1), (2),(4) ,(6),(7) and (10), thank you very much for your affirmations.

For (3), thank you very much for your advice. However, this paper mainly focuses on the improvement of the algorithm and the innovation of the data processing, and does not involve the whole system, so the system flow chart may not be applicable here.

For (5), (8) and (9), thank you very much for your advice. The variables, format and grammar have been checked by ourselves again.

Reviewer 5 Report

Authors are advised to include more comparative analyses of their proposed approach Fuzzy MLKNN algorithm.

Author Response

Response: Thank you very much for your advice. The comparative analysis has been supplemented in Section 4.3.

Round 2

Reviewer 3 Report

Thanks for taking the time to respond to my concern. I appreciate the author's response.

This literature review should include citations to the following papers:

https://doi.org/10.1007/978-3-030-24271-8_56

https://doi.org/10.48550/arXiv.1910.06588

https://doi.org/10.1007/978-3-319-90775-8_29

http://dx.doi.org/10.36785/jaes.111501