Visual Simulator for Mastering Fundamental Concepts of Machine Learning

Round 1

Reviewer 1 Report

This paper provides a good overview of existing courses in machine learning at the most famous universities in the world, as well as a description of the implemented visual software systems for learning these algorithms. The language used throughout this paper needs to be improved, the author should do some proofreading on it.

The aim of the research paper is very clear, and the topic corresponds to this special issue, so I recommend that the paper be accepted after these minor changes:

The authors devoted the first part of the paper to a detailed analysis of academic courses and created a very interesting categorization of software tools used in teaching. In tables 1 and 2, you should add a course from European universities (I think only Swiss universities are covered, and similar courses exist at universities in Ljubljana, Zagreb, Munich, Madrid, etc.). I recommend adding some information about that if you can find it publicly.

The importance of visual simulation in this area is not clearly highlighted, so explain why they are more important than standard textual problem solvers.

Working with such an application is not explained to beginners. How complicated is the installation of such a tool? What are the minimum platform requirements for this tool to work?

I am particularly interested in applying this tool to other datasets and extending it with other algorithms. Could it be applied in sports and in the prediction of physical activities? (see the paper: "An XGBoost-based physical fitness evaluation model using advanced feature selection and Bayesian hyper-parameter optimization for wearable running monitoring", Guo et al.)

How long does the proposed approach take to learn parameters (i.e. in described examples)? These details are missing and must be added to keep the paper complete. The limitations of the proposed study need to be discussed before or in the conclusion (one or two sentences are enough).

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

In this paper the authors presented a new tool to teach machine learning courses.  I do not think this is a right journal since this paper purely proposes a new tool for teaching not for research.  

I teach a lot of machine learning statistical learning courses over decades (and I get usually very high evaluation like higher than 4.7 out of 5, sometimes I get 5 out of 5).  So I know how to teach such courses.  I do not think it is difficult to teach programming but it is difficult to teach concepts behind of each package.  There are many easy software to use such as JMP.  But each student must know what they are doing.  Otherwise they cannot interpret output from the software.  Thus I do not agree with the authors that tools which make easy for students to code would help them to understand ML/AI materials more.  

The authors might want to tone down some texts.  For example 

line 24: I do not think AI consists of these two parts.  

line 25:  I do not think so.  If you mean gaming like Alpha go sure.  But you need to clarify.

line 202 and 203: But training is an important part of modeling.  

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The authors present an interesting work.  However, the following issues should be addressed:

1. The authors mention "The listed tools have been analyzed across multiple criteria, ...".  How the analysis of the tools was conducted? Questionnaires were provided? A specific scale was used? etc.

2. How the evaluation of the proposed tool was performed?

3. Qualitative and quantitative comparison of the proposed tool with the existing tools should be provided.

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

This is a review for the paper "Visual simulator for mastering fundamental concepts of machine learning" by Milakovic et al. I think it looks great but I
feel this paper should go to a journal for education not in this journal since this is for applications.  If the authors wish to publish this manuscript
I would suggest applications of their visualization method to empirical datasets in real life.

Comments for author File: Comments.docx

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 3

Reviewer 2 Report

I think it looks good enough for publication.