(Research project) Building a theory of data mining requires setting up a theoretical framework so that the major data mining functions can be explained under this framework. Take one theory as an example (e.g., data compression theory) and examine how the major data mining functions fit into this framework. If some functions do not fit well into the current theoretical framework, can you propose a way to extend the framework to explain these functions?