“…Feature location in models at the industrial scale is a central topic in previous works from our SVIT research group [16,17,18,19,20,21,22,23,24]. Given a feature description as input, these works [16,17,18] rank the model fragments that are relevant for the feature and explore different approaches to guide the automated feature localization: clustering (through Formal Concept Analysis) [16], empirical learning (through Learning to Rank) [17], and combinations of Similitude, Understandability, and Timing (through Latent Semantic Indexing, Model Size, and Defect Principle, respectively) [18].…”