Citation Context

... al. [21] use LDA to predict the best potential answerers to a new question; their LDA models capture different topics an expert is interested in. In the domain of software engineering, Hindle et al. =-=[22]-=- cluster commit-log comments and others cluster source code [23], [24], [25], all with LDA. On the contrary, our LDA model captures topics of mailing lists. The concept of applying data-mining techniq...

Citation Context

...ir use is built upon assumptions, such as usability, that have not been validated with stakeholders, such as developers. Often these topics are interpreted solely by the researchers themselves (e.g., =-=[12]-=-). Thus the use of topic analysis, LDA and LSI, and topics in software 978-1-4673-2312-3/12/$31.00 c○ 2012 IEEETopic Topic Requirements Topics Commits Figure 1. Our data model, Topics are extracted f...

Citation Context

...fect of various software engineering contexts on the accuracy of duplicate bug-report detection. These word lists include: Android architectural words [10], software Non-Functional Requirements words =-=[11]-=-, Android topic words extracted applying Latent Dirichlet Allocation (LDA) method [12], Android topic words extracted applying Labeled-LDA method [12], and random English dictionary words (as a contro...

Citation Context

...ructural information query can be used to find other related code or determine the impact of a proposed change. There is also work on automatically extracting topic words and phrases from source code =-=[28]-=-, [29], [30], displaying search results in a concept lattice of keywords [12], and clustering program elements that share similar phrases [31]. These techniques could be used in conjunction with phras...

Citation Context

...relations are established using the WordNet ontology [12]. Although we are aware that WordNet may not necessarily be the most suitable ontology to analyze source code, as pointed out by Hindle et al. =-=[13]-=-—at the moment WordNet represents, to the best of our knowledge, the most suitable technology for this task. Also, the approach is perfectly applicable even if replacing WordNet with a domain-specific...

Citation Context

... a range, or a number requires manual inspection, since it is mostly done through the based on the input data size, but they also suggest sensitivity use of heuristics [9], [11], [56]. analysis [17], =-=[29]-=-, [59]. However, manual inspection of every single output is not ✓ ✏ feasible. Therefore, sampling and verification approaches are S2: Carefully tuning parameters and performing sensisuggested by rese...

Citation Context

...nalysis techniques such as statistical translation both being trained on such diverged code bases, and also being used to merge them permanently or on-demand. Other techniques, such as topic analysis =-=[24]-=-, may help with categorization and a search for relevant functionality. We will see more procedures and tools to support various branching strategies [48], both within an enterprise, among companies, ...

Citation Context

...Many topics that are local to a small portion of the project’s life turn out to be relevant and interesting. Additionally, some topics recur over the life of a project. The authors later demonstrated =-=[17]-=- how many of these topics can be named using an automated approach. Lukins et al. [18] used LDA and Latent Semantic Indexing (LSI), a well-known information retrieval technique, in a study comparing p...

Citation Context

...ther tended to produce the best results. Typically steeped in the information retrieval domain, Poshyvanyk et al. [21, 18, 16] have explored the use of IR techniques such as LSI [15] and LDA to recover software traceability links in source code and other documents. For a general literature survey related to traceability techniques (including LDA), the interested reader should refer to De Lucia [7]. Baldi et al. [3] labelled LDA extracted topics and compared them to aspects in software development. Baldi claims that some topics do map to aspects, this was somewhat corroborated by Hindle et al. [12]. 16 Clustering is frequently used to compare and identify (dis)similar documents and code, or to quantify the overlap between two sets of documents. Clustering algorithms can potentially be applied to topic probability vectors produced by LDA. LDA has been used in a clustering context, for issue report querying and deduplication. Lukins et al. [14] applied LDA topic analysis to issue reports, leveraging LDA inference to infer if queries, topics, and issue reports were related to each other. Alipour et al. [1] leveraged LDA topics to add context to deduplicate issue reports and found that LDA ...

Citation Context

... al. [21] use LDA to predict the best potential answerers to a new question; their LDA models capture different topics an expert is interested in. In the domain of software engineering, Hindle et al. =-=[22]-=- cluster commit-log comments and others cluster source code [23], [24], [25], all with LDA. On the contrary, our LDA model captures topics of mailing lists. The concept of applying data-mining techniq...