timm
commented
7 years ago looks about half done. thanks for working on this
random comments, below
dont dig your title cant we go back to "trends in topics in SE (1993 to 2016)"?
IST requires abstracts are structured abstracts. Excatly the following headings. and there might be a word limit (for abstracts). plz check.
Your anstract need to start with a "Context" that sets up a problem. what is in your intro that sets up a problem; e.g. "Existing methods for studying large text corpuses can be (very) labor intensive and, systematic literature reviews and manual content analysis rely extensively on subjective human judgment. What is required is some method to define a literature analysis methods that is automatic, scalable insightful, and reproducible."
some echoes of previous paper persit. result2 should say "2 papers" not "3" . and is result3 really the current summary?
ditto with result4
let table1 and table2 take over the page. (remove the [!t] from \begin{figure}
table captions p3 v. small. did we do something evil in the preamble to shrink that text?
result7 p3 renders badly. result7 data is missing from this paper
table4 is 2009 to 2016. why that period? that's 7 years... why not 5 years
your entry for defects i table 4 is missing a new line.
anyway we can make the last line in table4 be 1 line?
when we introduce the topics in table2 can we do any sort of clustering to show their similarities? something like the hierarchy shown left hand side of fig3? then adjust the id of each topic to reflect those similarities. e,g, if fig3 lhs was the hierachy them i'd be tempted to say
old = new
1,9 = a1,a2
2 = b
0,10=c1,c2,
5,6 = d1,d2
3,4,7,8 = e1,e2,e3,e4
and just to say something u already know, those labels are not on the figures.
somewhere you are going to have to appease the Systematic literature review people. e.g. one approach is to make more use of people in some extensive and structures literature review. Here some text from the thing with Zhe that might help:
Previously we have analyzed the costs of manual systematic literature reviews SLRs [5, 6]. One part of that process, primary study selection, is one of the most difficult as well as time-consuming aspects in SLR. Usually, reviewers need to evaluate thousands of studies trying to find dozens of them that are relevant to the research questions based on their title, abstract, or full text. An extreme example of this is where reviewers sourced over 3,000 studies, and only used 7 of them in their final review [8]. The cost associated with primary study selection has become a serious problem and will continue to grow in the near future as the population of candidates for primary studies increases exponentially.
We do not intend the methods of this paper to replace manual SLRs. Rather, we propose a combination of techniques. When time is very limited and some summary of the literature is required in just a few days, we recommend the LDA approach of this paper. However, when resources permit and the deadline for the work is several weeks/months away. then we recommend manual SLRs.
[5] E.Hassler,J.C.Carver,N.A.Kraft,D.Hale,Outcomesofacommunity workshop to identify and rank barriers to the systematic literature review process, in: Proceedings of the 18th International Conference on Evalua- tion and Assessment in Software Engineering, ACM, 2014, p. 31. [6] J. C. Carver, E. Hassler, E. Hernandes, N. A. Kraft, Identifying barriers to the systematic literature review process, in: 2013 ACM/IEEE Interna- tional Symposium on Empirical Software Engineering and Measurement, IEEE, 2013, pp. 203–212.
[8] Y. M. Bezerra, T. A. B. Pereira, G. E. da Silveira, A systematic review of software product lines applied to mobile middleware, in: Information Technology: New Generations, 2009. ITNG’09. Sixth International Con- ference on, IEEE, 2009, pp. 1024–1029.
bigfatnoob
Citemap Paper
Classification