An architecture for continuous deployment of the Smart Collaborative Learning Service (SCLS) based on a predictive model to build complementary teams

Abstract

Abstract In this article, we tried to present a novel opportunity of sharing knowledge between teams of learners based on the completeness of their skills. We then, produced an intelligent collaborative learning service for the prediction of complementary teams that participate in a pedagogical method of sharing and acquiring knowledge within the academic environment. This smart service encapsulates a predictive model that is selected dynamically basing on its higher accuracy to achieve a continuous and high-performance result concerning team’s composition. Precisely, we covered the Smart Collaborative Learning service deployment in a cloud architecture (Amazon Web Service) to provide continuous and on-demand building complementary team service through their REST API interface. This experimentation led us to design a new architecture allowing the continuous deployment of a predictive model in a cloud environment while using streaming and integrated deployment architectures. We proposed a novel cloud architecture for continuously deploying a Smart Collaborative Learning Service that encapsulates a predictive model for building complementary teams of learners. Invocating this smart service in a collaborative learning platform creates a flexible building team of learners, based on the completeness of learner skills required by a collaborative project. This intelligent service meets the intelligence criteria that we cited in our approach to assess its ability to continuously respond to environmental changes and data updates, to effectively predict complementary teams. In the present work, we successfully deployed the intelligent collaborative learning service and make it powerful and on-demand sharing using its REST API. Thanks to the proposed architecture which includes intelligent layers from the semantic representation of the data, heuristic preprocessing, the development of predictive models and the streaming of data updates.