Association for Computing Machinery
Data cleaning aims at converting erroneous, incomplete, duplicate and inconsistent data into high-quality data. There is an extensive list of commercial data cleaning tools available and a significant number of data cleaning projects reported in the literature. Typically, a data cleaning process is modeled as a directed-graph of data transformations, which incorporate data cleaning criteria that must be satisfied by a data flow graph. Some data is rejected by a data transformation because it does not satisfy underlying criteria. One way to resolve these situations is to apply data correction procedures to the data flow graph.