Review Reports

Round 1

Reviewer 1 Report

The paper explores the new domain of social media and a research-based survey on the detection of fake news. 

The emphasis is on Content-based, Propagation-based and source-based. 

Although the initiative is good, however, the following points need to be addressed. 

1. Table-based summaries are provided, and statistical analysis should be explored

2. Backend algorithms used in the rating levels of common fact-checking websites need to be vetted and included in the survey. 

3. Detailed research-based survey needs to be explored apart from Content-based, Propagation-based, and source-based

Comments for author File: Comments.pdf

Can be improved. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

 

The paper is interesting but there are some issues that should be addressed in my opinion.

- in Fig. 3 it is not clear to me what the big arrow represents in the middle of the figure

- same consideration with the big double arrow in Figure 4

- In section 3.1 authors propose a learning function that is never used in the rest of the paper.

- what is really missing in this paper is a comparison between the various approaches to detect the fake news. This comparison should be made by evaluating which approach is better by, for example, counting how many times the learning function is working correctly.

- In other words, giving the same datasets and having information about which is fake news and which is not, what is the score of the various approaches proposed in Table 3?

- I am not sure if this kind of comparison can be done, but I would like to read some considerations regarding the efficiency of the various approaches. I think it would be interesting to just have a comparison between a small set of the approaches that are able to work on the same dataset.

I think is quite good.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

This paper is a survey of the literature. The authors focus on the proliferation of fake news through social networks and the subsequent challenges in detecting it. They highlight the significant attention this issue has garnered in computer science. While progress has been made in fake news detection, several hurdles persist. The paper aims to provide a comprehensive review, delving into various aspects of fake news detection, such as fundamental theory, feature types, detection techniques, and approaches. They propose a classification method after extensive research and organization. Additionally, the paper compares and analyzes datasets used in fake news detection across different domains. The inclusion of tables and visuals aids in a clearer understanding of the overall landscape of fake news detection.

 

In my view, the literature is well covered and the comparisons between the different results provide a good meta-analysis.

I believe that this paper will be a useful starting reference for the interested reader, who could come from many different disciplines, and any researcher focusing on this issue in the future.

 

If anything, I would have been interesting in knowing if there are differences between the topics of the fake news. E.g., are fake news on politics more or less easy to detect than those on climate change? Can the authors extrapolate any information about this from their sources.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors addressed all my comments and responded in a satisfactory manner. To my opinion manuscript has been substantially improved over its original version. Having that in mind, I advise to accept the paper.

Reviewer 2 Report

I'm still missing the experimental section. I see that there is some limitation but maybe some comparison could be done even with few method in a small dataset. In any case, this paper has a good "related work" discussion.