Towards Feature Location in Models through a Learning to Rank Approach

Utilizing Automatic Query Reformulations as Genetic Operations to Improve Feature Location in Software Models

“…In [6], the fitness function (the similarity to the feature description) is fixed and five search strategies are evaluated (Evolutionary Algorithm, Random Search, steepest Hill Climbing, Iterated Local Search with restarts, and a hybrid between Evolutionary Algorithm and Hill Climbing). In [85], a learning-to-rank approach is proposed to improve the fitness function to locate features in models, whereas the candidate model fragments are randomly generated. In [86], models at run-time are proposed to be used for increasing the information for feature location.…”

Definition of Descriptive and Diagnostic Measurements for Model Fragment Retrieval

“…Specifically, there is an emerging interest in leveraging machine learning techniques to address the challenges of MFL (D. Binkley and D. Lawrie 2014), (Corley, Damevski, and Kraft 2015), (B. Le et al 2016), (Ana C Marcén, Pérez, and Cetina 2017), (Ana C. Marcén et al 2017), (Mills, Escobar-Avila, and Haiduc 2018).…”

“…Finally, some of our previous works (J. Font et al 2017), (Jaime Font et al 2016, (Arcega, Jaime Font, Øystein Haugen, et al 2016), (Marcén et al 2017), (Lapeña Martí et al 2017 present Feature Location approaches to discover software artifacts that implement the feature in models.…”

“…Marcen et al (Marcén et al 2017) propose a feature location approach to discover model elements that implement the feature functionality in a model. Through a model and a feature description, model fragments that are extracted from the model and the feature description are encoded based on a domain ontology.…”

“…Le et al 2016),(Ye, Bunescu, andC. Liu 2014),(Ana C. Marcén et al 2017), (Binkley and Lawrie 2014). In these works, machine learning techniques take advantage of the knowledge and the experience generated for years in software companies to perform the feature location process.…”