This paper presents TD-AC which is an effective algorithm for the truth discovery problem when the attributes over data are structurally correlated. We build our procedure on an abstract representation of the truth in the data, the k-means clustering technique and the silhouette measure to automatically find an optimal partitioning of the input data (or a near-optimal) maximizing the accuracy of any base truth discovery process. The intensive experiments conducted on synthetic and real datasets show that TD-AC outperforms existing partitioning approaches with a more reasonable running time. It improves on synthetic datasets the accuracy of standard truth discovery algorithms by 6% at least and by 16% at most and also significantly when the data coverage rate is high for the other types of datasets
