On negative results when using sentiment analysis tools for software engineering research
Citations
179 citations
Cites background or methods from "On negative results when using sent..."
...[16] conducted a comparison of four widely used sentiment analysis tools: SentiStrength, NLTK, Stanford CoreNLP, and AlchemyAPI....
[...]
...in previous work [15, 16], is not sufficient to evaluate sentiment analysis tools....
[...]
...This “out-of-thebox” usage has been criticized due to the poor accuracy these tools achieved when applied in a context different from the one for which they have been designed and/or trained [16, 23, 35]....
[...]
...Previous literature has pointed our that sentiment analysis tools cannot be used outof-the-box for software engineering tasks [15, 16, 23, 35]....
[...]
...Given the warning raised by previous work in our field [16, 23, 35] there was the need for training and customizing the sentiment analysis tool to the Stack Overflow context....
[...]
155 citations
138 citations
Cites methods from "On negative results when using sent..."
...[89] On negative results when using sentiment analysis tools for software engineering research Laboratory Experiment Investigates to what extent results from SE studies using sentiment analysis depend on the choice of sentiment analysis tool....
[...]
115 citations
Cites background from "On negative results when using sent..."
...sentiment-analysis tools that are not trained specifically for the software engineering domain [29,30,42]....
[...]
...[29,30] conducted a classification study of seven datasets from technical websites (i....
[...]
114 citations
Cites result from "On negative results when using sent..."
...A similar approach is also used in other studies [28], [29] to categorize emotional expressions according to their polarities....
[...]
References
19,603 citations
"On negative results when using sent..." refers background or methods in this paper
...This requirement excludes such commercial tools as Lymbix3 Sentiment API of MeaningCloud4 or 2http://ow.ly/HvC5302N4oK 3http://www.lymbix.com/supportcenter/docs 4https://www.meaningcloud.com/developer/sentiment-analysis GetSentiment.5 Furthermore, we exclude tools that require training before they can be applied such as LibShortText (Yu et al. 2013) or sentiment analysis libraries of popular machine learning tools such as RapidMiner or Weka....
[...]
...A more careful sentiment analysis for software engineering texts is therefore needed: e.g., one might consider training more general purpose machine learning tools such as Weka (Hall et al. 2009) or RapidMiner15 on software engineering data....
[...]
..., one might consider training more general purpose machine learning tools such as Weka [35] or RapidMiner(15) on software engineering data....
[...]
...A similar approach has been recently taken by Panichella et al. (2015) that have used Weka to train a Naive Bayes classifier on 2090 App Store and Google Play review sentences....
[...]
16,435 citations
12,871 citations
"On negative results when using sent..." refers methods in this paper
...The second step uses the t-test or the rank-based Wilcoxon-Mann-Whitney test (Wilcoxon 1945), with correction for multiple comparisons, e.g., Bonferroni correction (Dunn 1961; Sheskin 2007)....
[...]
...The second step uses the t-test or the rank-based Wilcoxon-Mann-Whitney test (Wilcoxon 1945), with correction for multiple comparisons, e....
[...]
10,442 citations
7,604 citations
"On negative results when using sent..." refers background or methods in this paper
...Therefore, rather than reporting accuracy of the approaches we use the Weighted kappa (Cohen 1968) and the Adjusted Rand Index (ARI) (Hubert and Arabie 1985; Santos and Embrechts 2009)....
[...]
...One possible way to address this concern would be to associate different kinds of disagreement with different weights: this is an approach taken, e.g., by the weighted κ (Cohen 1968)....
[...]