Fuzzy Neural Network Expert System with an Improved Gini Index Random Forest-Based Feature Importance Measure Algorithm for Early Diagnosis of Breast Cancer in Saudi Arabia

Round 1

Reviewer 1 Report

Dear Authors, 

Here are a few changes I will suggest.

• Novelty is missing, please justify what is new, already a lot of studies being conducted on the same data, I may consider it as more a review paper than a research paper. Please change its title and add at least a groundbreaking literature review.
• Hypothesis and Research questions need to be discussed in details 
• The motivation of the study is extremely weak and must be improved.
• No statistical testing and its comparison alongside. Results can be compared without statistical confidence. 
• Discussion is weak and should be separate 
• Add the section Application of the machine learning in the literature review or in the introduction and connect with the motivation,  and referred a few multidisciplinary approaches of ML and cite some papers of them for eg, sentiment analysis, image processing etc. For your reference, a few of the references are provided below and if you find them relevant then cite them on the application of machine learning sections.
  • https://www.mdpi.com/2504-2289/3/3/37
  • https://www.mdpi.com/2504-2289/4/1/3
  • etc 

Thank you

Author Response

Response to reviewer

1. Novelty is missing, please justify what is new, already a lot of studies being conducted on the same data, I may consider it as more a review paper than a research paper. Please change its title and add at least a groundbreaking literature review.

The title of the paper was changed from “An Improved Gini Index Random Forest-Based Feature Importance Measure Algorithm Inspired Fuzzy Neural Network-Based Expert System Framework for Early Prediction of Breast Cancer in Saudi Arabia” to “Fuzzy Neural Network Expert System with an Improved Gini Index Random Forest-Based Feature Importance Measure Algorithm for Early Prediction of Breast Cancer in Saudi Arabia” as indicated in red in the manuscript.

1. Hypothesis and Research questions need to be discussed in details 

The Research questions were spelt out and discussed in detail as it can be seen red at section 1 of the manuscript

1. The motivation of the study is extremely weak and must be improved.

The motivation of the study has be improve, as it can be seen red at end of the section 1 of the manuscript

1. No statistical testing and its comparison alongside. Results can be compared without statistical confidence. 

We used confusion metric which is most appropriate statistical tools for predictive or classification models or system. The tool helps us to determine the AccuracySensitivity and Specificity of the expert system.

1. Discussion is weak and should be separate 

Section 4: is the result and discussion of the study and it is named “Fuzzy Neural Network Expert System with an Improved Gini Index Random Forest-Based Feature Importance Measure Algorithm”. It has been further improved as indicated in red.

1. Add the section Application of the machine learning in the literature review or in the introduction and connect with the motivation, and referred a few multidisciplinary approaches of ML and cite some papers of them for eg, sentiment analysis, image processing etc. For your reference, a few of the references are provided below and if you find them relevant then cite them on the application of machine learning sections.
1. https://www.mdpi.com/2504-2289/3/3/37
2. https://www.mdpi.com/2504-2289/4/1/3

We find the above paper very interested and it helps us to further improve our manuscript, as such we have cited the paper.

Reviewer 2 Report

The proposes Improved Gini Index Random Forest-Based Feature Importance Measure Algorithm Inspired Fuzzy Neural Network Based Expert System Framework  algorithm for early prediction of breast cancer. This work looks interesting; however, the following issues should be fixed to get accepted in this journal.

1. The title of this paper is vague. Please make it simple and easy to understand
2. The abstract needs to be rewritten because it has grammatical errors and lacks motivation and coherence.
3. Motivation is still unclear in introduction and related works
4. Please discuss about other feature extraction methods of breast cancer images as in: -Sitaula, Chiranjibi, and Sunil Aryal. "Fusion of whole and part features for the classification of histopathological image of breast tissue." Health Information Science and Systems 8.1 (2020): 1-12.

Author Response

Response to reviewer

1. The title of this paper is vague. Please make it simple and easy to understand

The title of the paper was changed from “An Improved Gini Index Random Forest-Based Feature Importance Measure Algorithm Inspired Fuzzy Neural Network-Based Expert System Framework for Early Prediction of Breast Cancer in Saudi Arabia” to “Fuzzy Neural Network Expert System with an Improved Gini Index Random Forest-Based Feature Importance Measure Algorithm for Early Prediction of Breast Cancer in Saudi Arabia” as indicated in red in the manuscript.

1. The abstract needs to be rewritten because it has grammatical errors and lacks motivation and coherence.

The abstract was rewritten as suggested and changes made, were in red color.

1. Motivation is still unclear in introduction and related works

We have revised both introduction and related works for the clarity of the motivation of the study. The changes made were in red color.

1. Please discuss about other feature extraction methods of breast cancer images as in: -Sitaula, Chiranjibi, and Sunil Aryal. "Fusion of whole and part features for the classification of histopathological image of breast tissue." Health Information Science and Systems 8.1 (2020): 1-12.

The feature selection method, we used in our manuscript is for text dataset. Therefore, they can’t be applicable to our research, because the feature extraction methods discussed in the above paper were for image dataset. However, we find the paper very interested and it helps us to further improve our manuscript, as such we have cited the paper.

Round 2

Reviewer 1 Report

Dear Authors,

I appreciated your improved version.    

Please consider statistical testing such as the T-test, Ztest and CHI test.  Confusion metrics should not be considered for statistical testing.

Check for typos.

Please provide a detailed explanation for RQs answers.

Thanks 

Author Response

Dear Prof.,

Thank you, very much for good review that makes our manuscript rich.

Thanks.

Ebrahem A. Algehyne et al.

Author's Reply to Reviewer

1. Please consider statistical testing such as the T-test, Ztest and CHI test.  Confusion metrics should not be considered for statistical testing.

Thank you for good observation and we have learnt a lot from it. We have performed the Z test as it can be seen in Section 6.

2. Check for typos.

We have checked for the typos/

3. Please provide a detailed explanation for RQs answers.

We have provided a detailed explanation for RQs answers as indicated in red at the end of the Section 5 of the manuscript.

Thank you, very much for good review that makes our manuscript rich.

Reviewer 2 Report

Accepted. Please check grammatical issues and add hyper-parameters of ML methods.

Author Response

Dear Prof.,

Thank you, very much for good review that makes our manuscript rich.

Thanks.

Ebrahem A. Algehyne et al.

Author's Reply to Reviewer

Please check grammatical issues and add hyper-parameters of ML methods.

We have checked grammatical issues and added hyper-parameters of ML methods as indicated in red at section 3.1.1 of the manuscript.

Thank you, very much for good review that makes our manuscript rich.