Machine Learning and Fuzzy Logic in Electronics: Applying Intelligence in Practice

Round 1

Reviewer 1 Report

Dear Authors

• Please organize the information in a better way, group it according to the content and objective of each variant. There are too many loose paragraphs. 
• Improve the graphics, they are too small and are not clear. 
• In the Thematic Evolution and Trending Topics, there is a lack of information, this concept appears in the '60s and not in 2010.

Author Response

Dear Editor,

Thank you so much for your comments and suggestions for improvements! We are very grateful! We performed the following corrections:

• The information is organized in better way: in sections 3 and 4 are formed 5 different subsections: 1. Scientific production, 2. Characteristics of sources, 3. Trending research topics, 4. Thematic evolution, 5. Analysis of selected papers.
• The graphics' size is improved in order to the text to be more clear. Several figures are removed to not doubling the explanations.
• We checked the figures with thematic evolution and trending topics. Sometimes we choose to visualize smaller investigated period (for example from 2014 to 2020 instead from 2010 to 2020) to not overload the graphics that could lead to their unreadability.

Thank you again for comments and suggestions!

Reviewer 2 Report

1. Too many pictures are not clear and overlap with text. (Fig. 7, Fig. 8, Fig. 20, Fig. 21)

2. The authors take the bibliographic data from indexing scientific databases Scopus and Web of Science. By the keyword: machine learning and fuzzy logic, there are only 434 and 199 documents identified in Web of Science. Most of the analysis is only based on Scopus. Why did you choose Web of Science?

3. Page 2, line 113-119: The authors use parameters MCP and SCP to indicate that the corresponding authors with the biggest contribution to the explored topic(Fig. 2). Please explain the significance of these parameters here and why they are not used in the fuzzy logic paragraph.

4. The authors use a bibliometric approach to map scientific knowledge regarding examined topics in a certain domain and use this method to analyze machine learning and fuzzy logic respectively. I think this method can be used in various fields, please explain the relevance of this method to machine learning and fuzzy logic.

5. To the authors’ best knowledge, the main theoretical contribution proposed in this work is the systematic picture. However, Bradford’s Law, h-index, three-fields plot, and thematic map are all methods that have already been proposed. Can you describe the benefits and contributions of your method in more detail?

6. Page 13, line 462-463: I can't find any connection between Figure 2 and annual scientific production, please explain it.

7. Page 13, line 460-463: The themes evolution is divided into time slices. Please explain the reason for this and why they are not used in the fuzzy logic paragraph.

8. Page 19, line 597-598: Please add relevant information about the h index of the Web of Science and explain why the highest impact according to database Web of Science is scientific journals.

9. In the thematic map part(Fig. 10-13), Please have a more detailed description of the changes in each time slice for the cluster, instead of simply describing the information on the map.

Author Response

Dear Editor,

Thank you so much for your comments and suggestions for improvement! We are very grateful! We performed the following corrections:

1. The pictures' size is improved to be more clear. Several figures are removed: Fig. 6, Fig. 8, Fig. 17, Fig. 20 and Fig. 21. Fig. 7 is improved through removing unreadable text - we think that the curve form of the Bradford's law is better to be shown.
2. It is correct that the number of papers indexed in Web of Science is smaller than the papers included in Scopus related to the investigated domain. However, we think that the information gathered from Web of Science could complement and clarify the conceptualization.
3. A short explanation regarding the parameters MCP and SCP is added in the fuzzy logic section too.
4. It is correct that the bibliometric method could be applied to examination of any domain. We think that it could be used to outline the "big picture" in the researched domain too, which together with analysis of literature could create more complex presentation of the conceptual framework.
5. We use all these tools (Bradford’s Law, h-index, three-fields plot, and thematic map) to understand the contemporary situation in the explored domain and to draw the conceptual framework. Our contribution is related to summarization and analysis the application of machine learning and fuzzy logic in electronics, which is done through usage of bibliometric method and systematic review.
6. It is our mistake and excuse us for that - the curve of Fig. 3 is used for forming several time periods. We replaced in the text Figure 2 with Figure 3.
7. The time slices in the thematic evolution are chosen according to the form of curve on Figure 3 that concerns different periods in scientific production (in machine learning section). We can repeat the same approach for explanation the evolution in the fuzzy logic area, but we decide to apply different approach through construction thematic maps. In this way, we show different approaches for analysis.
8. We reconstructed Table 1 and Table 4, including combined information from Scopus and Web of Science and ordering sources according to h-index. The analysis shows that the papers indexed in Web of Science are mainly published in journals (in our investigated domain).
9. Some explanations are added to the Fig. 10-13.

Dear Editor, we would like to thank you again for your comments and suggestions! We would like to inform you that we also performed other changes to take into account the comments of all reviewers. 

Reviewer 3 Report

The submitted manuscript is overall not well organized. Some critical information has been omitted that decreases the readability of this paper. Please see my comments for details.

Strength:

1. Broad literature concerning machine learning and fuzzy logic applications for solving electronics problems is surveyed. Specifically, the applied research methodology includes a bibliographic approach and detailed examinations of 66 selected papers. 

weakness:

1. Even though many prior works are surveyed, the manuscript lacks detailed explanations of representative papers in each research direction.

2. The statistic analysis reveals that machine learning and fuzzy logic techniques have increased interest over the last ten years in modeling different hardware-based intelligent systems. The conclusion is sound, but more nontrivial findings are expected. Besides, the future research directions are not clearly described, which should be completed in the next revision.

Author Response

Dear Editor, Thank you so much for comments and suggestions! We are very grateful! We performed several corrections:

1. The paper is re-organized as sections are divided in subsections.
2. A discussion is added after papers' analysis in each section.
3. The conclusion is improved and some points for future work are outlined.

Thank you again for comments and suggestions!

Round 2

Reviewer 1 Report

Dear Authors

Now is much better, congrats. 

Reviewer 3 Report

Generally, I am satisfied with the revision. Please remove all the highlights in the final submission.