New Configurations for the Correction of the RDF knowledge bases

In each RDF knowledge base, several errors must be corrected by correction methods. Correction methods can be divided into three classes for the correction of outliers, inconsistencies, and erroneous relations. RDF knowledge base outliers can be considered as two types of outlier entities and triples. Inconsistent triples are corrected by inconsistency correction methods and there are many erroneous relation correction methods that each of them is used for a special objective. The variety of these errors is so wide so that no correction method could be able to cover them all. Most of the correction methods have been focused only on some of these errors, so a comprehensive study is mandatory to cover all of these elements for different objectives. Nevertheless, a couple of survey articles on the RDF knowledge base correction exist, but they are out-dated and did not present different configurations of these errors for various objectives. Since there is no configuration in this field, a new general configuration of the RDF knowledge base correction for a different objective is proposed here that can cover these various errors. In this configuration, a new classification of the errors is presented in which they are divided into three classes. The correction of each class is performed in a separate step. Finally, the state-of-the-art approach of each step is identified for each objective and a different configuration of these methods will be proposed for various objectives.