Citation Context

... to the feature location problem because developers usually construct queries without precisely knowing the vocabulary of the target system. Second, in our previous work on traceability link recovery =-=[30]-=-, we showed that LSI outperforms VSM and Bayes classifiers. Internal source code documentation has been recognized as commonly written in a subset of English [31] that lends itself to the use of infor...

Citation Context

...ware engineering problems, related to concept location, which have been addressed using LSI are concept [1] and feature [16] location, traceability link recovery between source code and documentation =-=[24, 25]-=-, tracing requirements [26] and other software artifacts [27], etc. Details on how LSI is used for feature location in source code are available in [1] and [13]. 4. Concept Location using Concept Latt...

Citation Context

...lated to feature location, such as software reuse [14, 19, 39], abstract data types identification [20], high level concept clone detection [21], traceability link recovery between software artifacts =-=[2, 6, 22]-=-, topic identification in source code [18], requirements tracing [16], etc. We introduced previously [24] a methodology to index and search the source code using LSI. The methodology was subsequently ...

Citation Context

...facts are represented in different formats and textual information is the common denominator for all of them, hence IR methods have become widely used in tools that support traceability link recovery =-=[1, 7, 11, 15, 18]-=- and in many recommendation systems [5]. In impact analysis [2, 4] the textual information indexed with the IR method is usually used in conjunction with structural information or historical informati...

Citation Context

...licit connection of software artifacts. Promising results have been achieved using Information Retrieval (IR) techniques [3], [4] to recover links between different types of artifacts (see e.g., [1], =-=[5]-=-). IR-based methods propose a list of candidate traceability links on the basis of the textual similarity between the text contained in the software artifacts. The conjecture is that two artifacts hav...

Citation Context

...dly be underestimated. Identifiers and comments represent an important source of domain information that is used by (semi-) automated techniques to recover traceability links among software artifacts =-=[1, 24]-=- and locate features in source code [8, 21, 25, 27, 29, 30]. Prior work [16, 32] employed a natural language-based representation of source code, based on the conjecture that there is an intrinsic pat...

Citation Context

...ceability link recovery. Several approaches have been developed to recover traceability links between source code and external documentation using probabilistic IR, vector space models [1, 9] and LSI =-=[7, 9, 17]-=-. Other work proposed a set of approaches to recover traceability links among requirements [9], requirements and source code [1, 9, 17], and requirements and test cases [15]. A set of tools that integ...

Citation Context

...c network models as similarity metrics in semi-automated links recovery. They found both methods promising. Subsequent work expanded on the IR paradigm that Antoniol et al. established. Marcus et al. =-=[17]-=- in particular expanded on the use of Vector Space Models (VSM) for traceability link recovery, using LSI; they claimed to obtain results at least as good as Antoniol et al. with less preprocessing re...

Citation Context

...st strengths) and avoid things thatschallenge analysts (or handle these challenges).sRecall and precision are measures frequently used tosevaluate the accuracy of a TM from a researcher’ssperspective =-=[8, 9, 10, 11, 12]-=-, while measures such as lag,sselectivity, and mean average precision have been used tosevaluate the quality of a TM from an analyst’s perspectives[5, 13]. In general, automated methods return candida...