Entity-level Sentiment Analysis of Issue Comments

1. Does the paper propose a new opinion mining approach?

Yes

2. Which opinion mining techniques are used (list all of them, clearly stating their name/reference)?

SentiSW, introduced in the current paper; SentiStrength-SE [1] and SentiStrength [2]. [1] Md Rakibul Islam and Minhaz F Zibran. 2017. Leveraging automated sentiment analysis in software engineering. In Proceedings of the 14th International Conference on Mining Software Repositories. IEEE Press, 203–214. [2] Mike Thelwall, Kevan Buckley, Georgios Paltoglou, Di Cai, and Arvid Kappas. 2010. Sentiment strength detection in short informal text. Journal of the Association for Information Science and Technology 61, 12 (2010), 2544–2558.

3. Which opinion mining approaches in the paper are publicly available? Write down their name and links. If no approach is publicly available, leave it blank or None.

SentiSW https://github.com/Jasmine-DJ-420/SentiSW SentiStrength-SE https://laser.cs.uno.edu/resources/ProjectData/SentiStrength-SE_v1.5.zip SentiStrength http://sentistrength.wlv.ac.uk/

4. What is the main goal of the whole study?

Introduction of a new method combining sentiment analysis and entity recognition

5. What the researchers want to achieve by applying the technique(s) (e.g., calculate the sentiment polarity of app reviews)?

As above, introduction of a new method combining sentiment analysis and entity recognition

6. Which dataset(s) the technique is applied on?

SES: 300 comments sampled from 231,732 issue comments from 10 popular repositories over 10 different main languages from GitHub such that i) each project contains at least 5,000 stars; and ii) each project contains at least 5,000 issue comments JIRA issue comment set of Ortu et al. Marco Ortu, Alessandro Murgia, Giuseppe Destefanis, Parastou Tourani, Roberto Tonelli, Michele Marchesi, and Bram Adams. 2016. The emotional side of software developers in JIRA. In Mining Software Repositories (MSR), 2016 IEEE/ACM 13th Working Conference on. IEEE, 480–483.

7. Is/Are the dataset(s) publicly available online? If yes, please indicate their name and links.

Yes. SES https://github.com/Jasmine-DJ-420/SentiSW/tree/master/data Ortu dataset http://ansymore.uantwerpen.be/system/files/uploads/artefacts/alessandro/MSR16/archive3.zip

8. Is the application context (dataset or application domain) different from that for which the technique was originally designed?

N/A

9. Is the performance (precision, recall, run-time, etc.) of the technique verified? If yes, how did they verify it and what are the results?

On SES: precision, recall, F1 are reported for different ML algorithms. In general, for the neutral class the tools achieves good precision and recall (82-88%); for the positive class precision (83%) is clearly better than recall (65%), for the negative class both precision and recall are low (40%). On the JIRA comments set accuracy of 87.82% is reported.

10. Does the paper replicate the results of previous work? If yes, leave a summary of the findings (confirm/partially confirms/contradicts).

No

11. What success metrics are used?

N/A

12. Write down any other comments/notes here.

The tool is reported to outperform SentiStrength and SentiStrength-SE "overall", as well as on precision of the positive class and F1, as well as on the recall of the neutral class and F1. For the negative class the performance is comparable to SentiStrength-SE.