Recommender systems help users to get in contact with relevant items for example books, products, people, from an overwhelming and complex information network. In software engineering, reusing software products or software by products is spreading and evolving. Good quality software domain models are valuable resources that can be reused in developing new products. Using recommender system to assist in smart and easy reuse of domain models is the motivating concept of this thesis work.
Among the enormous collection of models in model libraries, defining approaches and techniques to identify appropriate models pertaining to user?s interest is primarily addressed in this report. The fundamental and supporting concepts, a model recommender could incorporate to do intelligent recommendations are discussed in this work. This was achieved by tailoring concepts from contemporary recommender systems such as Amazon.com, to modelling domain. User?s contextual information such as details about his un?nished models and keywords he uses to search in library are analysed. These learnt speci?cations are used to extract useful elements from the model library. The proposed methodologies were evaluated theoretically and empirically to verify its requisites. The experimental results promises a research direction using our techniques and methods.