DATA CLEANSING SYSTEM, DATA CLEANSING METHOD AND PROGRAM FOR DATA CLEANSING

摘要

[Problem] In data cleansing using functional dependencies of related technology, a correction for which the correction cost is minimal is selected. In the case, when there are a plurality of correction tuple candidates for which the correction cost is the same, it is not possible to select the most probable among the same. [Solution] This data cleansing system is provided with a data rule set storage unit (100), an exception generation cause condition identification unit (110), a tuple correction unit (120), and a corrected data output unit (130). By estimating cause conditions which are common among a plurality of exceptions in order to operate so as to determine the most probable correction tuple candidate on the basis of the estimated cause conditions, it is possible to select a probable correction tuple candidate even in a situation in which a determination is not possible by using only the correction costs.