Secure Internet Financial Transactions: A Framework Integrating Multi-Factor Authentication and Machine Learning

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1.      Overall, I do not see what the novelty is here. Apart from usual authentication, an extra step of face recognition (not biometrical characteristics) is proposed. I cannot see how this could help in e-commerce security. This study lacks in many aspects. I am no sure how this could fit in an AI Journal, it is a common application, based on a common dataset used for testing relevant algorithms. Most parts could belong to an informative tutorial, not a scientific publication. I would only consider this for publication if a true, major revision is made, and I do not know if this is possible.

 

2.       “Because of that, this study presents a novel framework that combines multifactor authentication (MFA) and Machine Learning (ML) to overcome this issue.”

Rephrase the sentence – and decide when to use capitalized letters – if used in ML then they should be also on MFA (or not). And the whole abstract should be re-written.

 

3.      E-commerce – e-commerce

 

4.      “A new wave of opportunities for both consumers and cybercriminals has emerged as a result of the reliance on online services for banking and payments.”

This sentence is the same in the Introduction AND the abstract.

 

5.      person validates himself . Not herself?

 

6.       used example of this. which is not… Error here, too

 

7.      If multifactor authentication (MFA) is firstly referred in the text, then every time needed, refer to it as MFA only. The same for ML.

 

8.      Using the power of machine learning in classification problems will offer the highest possible security. At the same time, we will gain a user-friendly MFA implementation in which the user will authenticate himself only in two steps and other security layers may  be added depending on the ML classification. – Rephrase

 

9.      This paragraph should be better written in bullets, not like this “The significance of the study lies in its potential to …..”

 

10.   The introduction of new technologies that provide fast, cheap, and easy-to-use financial services; causes this growing adoption of financial services- rephrase

 

11.   The study of [33], proposed a multi – rephrase

 

12.   Sections 1 and 2 should be merged in one Introduction, keeping only the important things. There are many meaningless examples and hard-to-understand sentences.

 

13.   Figure 1 needs better description about the flow of information.

 

14.   Is Figure 3 original from the authors? Or from another source that is not mentioned?

 

15.   The authors claim: “Unfortunately, we are unable to offer the original characteristics and additional context for the data due to confidentiality concerns”. While they obtained the dataset from a public repository, free to all. Something goes wrong here….

 

16.   In general, reading this manuscript is difficult. There are fundamental English grammatical and syntactical errors. I cannot keep on correcting, the Discussion and the Conclusions Section need to be rewritten from scratch.

 

17.   The ML methods proposed are well-studied and defined in the literature. Moreover, these are not suggested for accurate/trustworthy applications as the one shown here (see, for example, ensemble and stacked methods that could be even more robust).

 

 

Comments on the Quality of English Language

Language quality is poor, difficult to read and understand.

Author Response

Many thanks for your comments and suggestions

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

In this study, an application has been developed to ensure the security of Internet financial transactions by applying multi-factor authentication and machine learning. The methodology employed in the study consists of three steps. Firstly, a credit card fraud dataset was obtained from the Kaggle website, and four different classifiers—Random Forest, Decision Trees, Logistic Regression, and Naïve Bayes—were trained. The accuracy of each classifier was analyzed, yielding percentages of 96.717%, 97.881%, 97.938%, and 92.354%, respectively.

Secondly, multi-factor authentication was performed, and thirdly, an e-commerce application interface was designed and implemented.

1. 1-The study is highly practical in this regard, focusing on financial applications. While the abstract and conclusion sections enumerate the actions taken in the study, the superiority of the study is not explicitly stated.

2. 2-The mathematical background of the study needs to be further developed.

3. 3-The discussion section should be more detailed. Particularly, comparisons with existing literature should be presented in tabular form, outlining the strengths of the current study.

4. 4-Figures 3 and 4 do not provide a clear visual representation and seem to lack meaningful content; I recommend their removal. The results in Table 1 should be elaborated on in detail.

Author Response

Many thanks for your constructive comments and suggestions

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This research presents a novel framework that integrates multi-factor authentication and machine learning to enhance the security of internet-based financial transactions. The methodology is based on using two layers of security, with the second layer activated if the machine learning model classifies the current transaction as fraud. The study uses four supervised machine learning classifiers, namely Random Forest, Decision Trees, Logistic Regression, and Naïve Bayes, to build a reliable machine learning model to address the credit card fraud problem. I would suggest revisions as listed below:

1.      How can the proposed framework be implemented and tailored to different E-commerce platforms and online payment systems?

2.      What are the potential challenges and limitations of integrating multi-factor authentication and machine learning in securing internet financial transactions?

3.      What are the user perceptions and experiences regarding the usability and effectiveness of the two-layer security approach in online financial transactions?

4.      The littrature review of this paper can be improved. Many recent researches using RF, Decision tree and Bayes methods to solve real-world challenges. Some references to build the related work

https://www.mdpi.com/2072-4292/14/13/3228

https://www.sciencedirect.com/science/article/abs/pii/S0306454922004613

 

 

Comments on the Quality of English Language

english is fine

Author Response

Many thanks for your constructive comments and suggestions

Author Response File: Author Response.docx

Reviewer 4 Report

Comments and Suggestions for Authors

In this paper, AlsharifHasan Aburbeian and Manuel Fernández-Veiga examined secure internet financial transactions through Multi-Factor Authentication (MFA) and Machine Learning (ML). This paper is particularly interesting and holds significance for the field of ML. It is well-written and well-structured. Based on this, I recommend publishing this paper after several modifications. Below are my suggestions:

 

Line 4: Please verify the names of the authors to ensure their correctness.

 

Line 20: "96.717%, 97.881%, 97.938%, and 92.354%" -> 96.717%, 97.881%, 97.938%, and 92.354%.

 

Lines 73-76: The following paragraph is suggested for deletion: "The remainder of the paper is as follows: - Section 2 discusses related studies in the field. Section 3 describes the methods and materials used to secure Internet financial transactions. The results of the study are shown in Section 4. Section 5 discusses the study results and indicates future directions. Section 6 draws the paper's overall conclusions." Authors may consider adding the main research target or contribution in this part.

 

Line 77: "Related Work" -> "Literature Review"

 

Line 108: "3.2. Related Studies" -> I suggest adding sub-titles with information like "3.2. Literature on Proposed Frameworks for Securing Transactions."

 

Line 241: "1.3. Machine Learning Phase" -> "3.3. Machine Learning Phase"

 

Line 252: "Intel(R) Core (TM) i5-10210U CPU @ 1.60GHz 2.11GHz" is repeated information from Line 348. Please check that related writing for the experiment is not repeated.

 

Line 273: "Table 1. Dataset Sample." I suggest making the caption self-contained, like "Table 1. Dataset Sample. V1 represents... Class 1 represents... Class 0 represents..."

 

Line 293: "Figure 4. The distribution of transaction type (non-fraud is zero, fraud is 1)" -> "Figure 4. The distribution of transaction type (non-fraud is 0, fraud is 1)"

 

Line 429: Table 2.

"support" -> "Support"

"96.717" -> "96.717%"

"97.881" -> "97.881%"

"97.938" -> "97.938%"

"92.354" -> "92.354%"

 

Comments on the Quality of English Language

Good quality on writing.

Author Response

Many thanks for your constructive comments and suggestions

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have re-written the paper and I am satisfied with the result. Now I believe it can be published in the Journal.

Comments on the Quality of English Language

There are still some minor language errors, maybe another pass would be beneficial.

Author Response

Many thanks

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This paper is acceptable as is.

Author Response

Many thanks

Author Response File: Author Response.docx