A Novel Framework for Identifying Customers’ Unmet Needs on Online Social Media Using Context Tree
Round 1
Reviewer 1 Report
The list in the introduction should be replaced by natural language.The list content is very simple and can be removed and replaced by related work.
The contributions of the study in the introduction are written in a fragmented way.
Suggestion 1: to have just one paragraph that explains the study contribution.
Suggestion 2: the outline of the report should be provided at the end of the introduction. Also, a brief description of why this solution is novel and how it differs from other methods.
Most references used in the LR (section 2) are more than 10 years old. As this topic related to sentiment analysis on social media, the review is obselete.
Suggestion 3: include a table that summarises the recent state of the art on sentiment analysis relating to the paper topic.
More importantly, the motivation for the idea is not well presented.
" This study defines a process that derives ideas to determine new functions or specifications in new products by utilizing social media" .
This is not clear. Suggestion 4: The problem being investigated is not clear. Why you need ideas? for what purpose? what is the aim of this proposed framework?
Figure 1 is a high level and should be explained in the text.
Section: 3.1. Determining, gathering and pre-processing the target social media data
reads like an essay/related work. What is the actual data collection method you used in the design of the framework? This needs to be clearly outlined (suggestion 5)
A UML diagram or some sort of a flowchart (e.g., Sequence or activity diagram) is required to outline the steps of the methodology.
Suggestion 6: more details are needed on the setup and technical details of the experiment. At the moment it simply looks like a scrapping tool which is widely available as commercialised software. More details on what was done with the scraped data are needed.
The conclusion is not well written. It goes straight into the limitation without firstly summarising the contribution of the framework.
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The authors provide very timely, pertinent, interesting and innovative research.
Regardless, the authors must provide further information (include more references and citations) to sustain their claims.
For instance:
a) the information in Table 1 is supported by which research(es)?;
b) lines 150-152, "For data collection, it is most preferable [sic] to use application programming interfaces (APIs) of corresponding sites." - according to whom?;
c) lines 159-160, "After data gathering, pre-processing [sic] should be conducted. For text data, Natural Language Processing (NLP) tools in accordance with a language type should be used." - please, explain why;
d) lines 159-184, a lot of information is conveyed, without a single citation;
e) The discussion of the results and, therefore, the conclusion, are very self-centred, i.e., the benefit in using the novel proposed tool by comparison to more traditional approaches is not clear. The authors should perform a more comprehensive discussion regarding that.
Some issues regarding English language and style must be addressed, for example:
a) line 142, "First, is determining data to analyze as a target among social media content." - the beginning of the sentence shall be rephrased. It is not clear which context the authors are bridging;
b) "most preferable" is redundant and shall be corrected;
c) line 288, shall be "(...) is not sufficiently (...)".
So, overall English changes are required.
Further on, it should be noted that the use of the first person should be avoided, as the use of the passive voice is preferable.
Line 293, an abbreviation is used, without being previously explained "UI".
More emphasis should be given to the presentation of the novelty tool - the context tree - which is a production from the authors.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
