Academic Editor: Youssef EL FOUTAYENI
Received |
Accepted |
Published |
25 September 2019 |
10 October 2019 |
10 March 2020 |
Abstract: Every company, considers documentation as the engine of its productivity. Recommendation systems allow the diffusion of knowledge between experts and operators in an efficient and a fast manner [1]. The goal of a trust-based recommendation system is to generate personalized recommendations by aggregating the opinions of users in the trust group [2]. In this paper, we present new modelling approach to build trust communities for recommending relevant documents to industrial operators based on collaborative filtering (CF) algorithm and Hierarchical Ascending Classification (HAC) method. The idea is that documents recommendation depends not only on operators’ needs and preferences but also on the trust relationships between these latter. Based on this idea, we suggest reliable documents recommendation tool in the industrial environment where operators can express their preferences towards a set of documents. An analysis of trust among operators within a group is performed to identify trusted operators. The groups of operators are formed by the HAC method. This approach is based mainly on the trust model that calculates the importance of a user within his group according to three metrics: his implicit self-confidence, his interests, and his activity score. This model allows a significant reduction of recommendation time by taking into account the group of trusted operators to predict new reliable recommendation.
