In the meantime of the rapidly growing of Linked Data, the quality of these datasets is yet a challenge. A close examination of the quality of this data could be very critical, especially if important researches or professional decisions depend on it. Nowadays, several Linked Data quality metrics have been proposed which cover numerous dimensions of Linked Data quality such as completeness, consistency, conciseness and interlinking. In this paper, we propose an approach to enhance the conciseness of linked datasets by discovering synonym predicates. This approach is based, in addition to a statistical analysis, on a deep semantic analysis of data and on learning algorithms. We argue that studying the meaning of predicates can help to improve the accuracy of results. A set of experiments are conducted on real-world datasets to evaluate the approach.