Classification problem in conditions of distorted cause-and-effect relationships

The paper considers the model for classifying an object stream for the presence or absence of a certain property $A$ in each object $O$. It is assumed that there are $M$ bijective transformations of objects coming for classification and in the stream, there are objects obtained from $O$ according to one of these transformations. For each object $O$, it is known that it contains property $A$ which causes the known objects $B_1, B_2, \ldots , B_k$ to appear in information spaces $I_1, I_2, \ldots , I_k$. This means that property $A$ can only be detected by observing the consequences $B_1, B_2, \ldots , B_k$. The problem is that for each object in the flow, it is necessary to determine the presence or absence of the converted reason $A$ in it. The algorithms for checking such possibility are built in cases where there is a description of characteristics $A$ and when such a description is absent.