R.ECOS - plataforma para suporte de um ecossistema de software para sistemas de recomendação

Abstract

Recommender Systems (RS) attempt to present relevant information to users at the time of consumption. The need to recommend resources in different application domains and the need to develop solutions focused on the reuse of RS components, creates an interesting scenario for adopting solutions from the Software Ecosystem (SECO) perspective. A SECO for RS should allow, in addition to interaction between actors and technology, integrations with others systems and platforms that support others SECO. By proposing a technology platform that supports a SECO, it is possible to assist researchers in understanding the different ways that organizations relate. By applying the SECO perspective in a specific domain, it is possible to centralize the requirements for developing solutions, facilitating the reuse, creation and evolution of specific techniques and approaches. The granularity of RS solutions, without the possibility of defining architecture patterns, combined with the benefits related to the reuse and sharing of approaches for SR, demonstrates the need to develop a technological platform where it is possible to create new solutions, taking advantage of existing solutions and encouraging research in the two main areas of this work, Recommendation Systems and Software Ecosystem. In this way, the problem addressed by this study is the integration of the methods, techniques and approaches of existing RS in a systematic and centralized way, being possible to facilitate the implementation of new solutions in RS, and also to promote the reuse and sharing of solutions and also the collaboration among the involved actors. The general objective of this study is to propose R.ECOS, a technological platform to support a software ecosystem to recommend resources, allowing the integration between their solutions and other ECOS. The evaluation of the proposal was carried out in two stages. First, two Feasibility Studies were defined to validate the used technology and the proposed architecture. Later, two Case Studies were carried out in a real context of use. The results indicate the feasibility of the study proposal.