Dear Evan Gertis

I suggest:

• Summarize the argument so that the writer can see/hear whether you have grasped what they think they have said.
• Identify the contribution that you think the article makes.
• Identify the strengths of the article.

Best Regards

Dr. Fatemeh Khozaei

beyond what Fatemeh Khozaei rightly said, I would add to stratify the results according to the quality of the articles, perhaps in two groups. Sometimes the reviewers ask for this

I think that first, you should focus on presenting and justifying your approach, paradigm and / or the theoretical categories that will be put into play for the development of the work.

Fatemeh Khozaei,

Thank you for your reply. I've formulated responses to your questions.

• Our hypothesis is that by removing the most common 1000 English words from 100 reddit posts we will be left with philosophical topics that describe the reddit community.

2. What are the contributions that each article makes?

• Exploring the Space of Topic Coherence Measures by Michael Röder and others.
• The authors of this paper present a framework that constructs existing word based coherence measures.
• Surveying a suite of algorithms that offer a solution to managing large document archives by David M. Blei
• The authors of this paper explain how topic modeling enables us to organize and summarize electronic archives at a scale that would be impossible by human annotation.
• Reading Tea Leaves: How Humans Interpret Topic Models by Jonathan Chang and others.
• The authors of this paper discuss machine probabilistic models for unsupervised analysis of text.

2.What are the strengths of each article?

• Exploring the Space of Topic Coherence Measures by Michael Röder and others.
• They describe an application of coherence measures used in TopicExplorer. They suggest that the application of this research would help improve topic quality by hiding topics that are irrelevant to a user.
• Surveying a suite of algorithms that offer a solution to managing large document archives by David M. Blei
• The example shown in the paper took 17,000 articles from science magazines and identified 100 topics. The application shown in this example shows that topic modeling provides an algorithmic solution to managing, organizing, and annotating large archives of texts.
• Reading Tea Leaves: How Humans Interpret Topic Models by Jonathan Chang and others.
• They focus on providing a predictive model of future text and a latent topic representation of the corpus. These studies show that they capture aspects of the model that are typically undetected by previous measures of models based on held-out likelihood.

Christian Spreafico,

In my opinion the stratification for the articles should follow:

1. Exploring the Space of Topic Coherence Measures by Michael Röder and others.

2.Surveying a suite of algorithms that offer a solution to managing large document archives by David M. Blei

3.Reading Tea Leaves: How Humans Interpret Topic Models by Jonathan Chang and others.

The reasoning behind this is that the first article explains what topic coherence is, the second article explains how we can use topic modeling to organize/view data, the third article explains how we can apply it. I'd love get your feedback on this.

Federico Del Giorgio Solfa,

I am presenting an example of applying topic modeling. The justification of this is to explanation for implementing topic modeling. As for theoretical categories, this research falls under the category of text mining. Does this sound coherent to you?

Thank you all for your feedback. I look forward to hearing from you.

Dear Evan Gertis

I agree with you on the procedure. My proposal was simply a stratification of the documentary sources to be analyzed but without changing the procedure

Hi Evan. I have reviewed and edited your article. The main changes required are: many spelling mistakes need correction, your opinion is omitted, you haven't commented on the third reference, and APA style is required for the references. You will see all the changes and comments in red. I hope this helps :)

Andrew Paul McKenzie Pegman Thank you for your help. I've removed the unnecessary parts of the paper.

Fatemeh Khozaei I've incorporated your feedback into the literature review.

Federico Del Giorgio Solfa Can you tell me what you think about the presentation and justification?

Evan Gertis ,

I would move the LDA description to a background section or algorithm description section were you state the algorithm plus improvements if any. I would start the introduction with a brief statement on topical models, their importance and state the lack of "papers that describe practical implementations of a topic modeling system". I would further expand on why the ' fields of philosophy' can contribute to text mining in the extraction of 'meaningful philosophical topics from conversations on reddit'

I found section 2 difficult to read. I would, for example:

weave the discussion as:

• initial author-algorithm name-brief description. Second author - algorithm name - why it is an improvement over the previous one, and so on
• Separate into a different paragraph the topic of the dataset since it does not have to do with the algorithms
• The text "This paper was used to develop the gensim library" is ambiguous since it is not clear if it is from a previous work or the current work.
• "The third paper uses three topic models to evaluate two models. They provide an in depth explanation of LDA analysis.". The pronoun "They" is confusing since the previous sentence is referring to the third paper yet you use 'they' instead if 'it'

As a further point why the focus on those references? The work by Boyd et al. while it is seminal there is much work done after that ( 2293 forward citations on Google Scholar). Is the system by Boyd still state of the art in 2021 or have the authors or other people improved the system? While I have not focused on LDA, these points came immediately to mind and more likely from a reviewer with in depth LDA analysis.

Regards

Fatemeh Khozaei

Can I get your feedback? I am stating what each paper presents, what their contribution was, and the strength of their argument is.

3 Reading Tea Leaves: How Humans Interpret Topic Models by JonathanChang and others

3.1 Authors

Jonathan Chang, Jordan Boyd-Graber, Sean Gerrish, Chong Wang, David M. Blei

3.2 Overview

The authors of this paper discuss machine probabilistic models for unsupervised analysis of text. They focus on providing a predictive model of future text and a latent topic representation of the corpus. These studies show that they capture aspects of the model that are typically undetected by previous measures of models based on held-out likelihood. The authors of this paper discuss machine probabilistic models for unsupervised analysis of text.

4 Exploring the Space of Topic Coherence Measures

4.1 Authors Andreas Both, Alexander Hinneburg

4.2 Overview

The authors of this paper present a framework that constructs existing word based coherence measures as well as new ones by combining elementary components. They present combinations of components which outperform existing measures with respect to correlation to human ratings. The authors suggest that results can be transferred to further applications in the context of text mining and information retrieval

5 Surveying a suite of algorithms that offer a solution to managing large document archives

5.1 Authors

David M. Blei

5.2 Overview

The authors of this paper explain how topic modeling enables us to organize and summarize electronic archives at a scale that would be impossible by human annotation. The example shown in the paper took 17,000 articles from science magazines and identified 100 topics. This paper describes how topic modeling provides an algorithmic solution to managing, organizing, and annotating large archives of texts.

I'd like to reword these paragraphs so that they flow together better. Any help would be greatly appreciated.

Here is a finally copy of my paper