Detecting mistakes in binary data tables

We suggest a classification of possible mistakes in lines of binary data tables (in formal contexts) and discuss their detection. An approach is proposed to detect some types of mistakes in new lines (contents of objects) of binary data tables (formal contexts). This approach is based on detecting those implications from implication bases of the formal contexts that are not met by a new object. It is noted that this approach can result in a complex computational solution. An alternative approach based on computing closures of subsets of object’s intent. This approach allows one to find a polynomial algorithm for the solution. The algorithm of unnecessary and extra missing properties can also be used. The results of experiments are dealt with.