In recent years, several education systems have been developed. Consequently, each learner can have different profiles which each one is related to a system. Each profile can be completed and enriched by the data coming from the other profiles in order to return results reflecting the learner's need. The profile enrichment requires the establishment of an interoperable system which i) resolves the problem of learner's profile heterogeneity based on a matching process and ii) integrates the data in the different profiles based on a data fusion process. The data fusion approaches mainly aim at resolving the conficts occurring in the data values. They are based on non organized proles which may produce inconsistent results. The profile organization is done either by using the machine learning techniques or the notion of temperature. In this paper, we propose a new data fusion approach to improve the conflict resolution by organized profiles. Each profile is organized by respectively merging a clustering algorithm and the temperature and by taking into account the data semantic relationship.