What is KDD?

The term KDD was probably first introduced in 1989 at a workshop in Detriot, Michigan.The etymology of the term may be further determined by dissecting its components. Knowledge refers to "the condition of comprehending truth or fact through reasoning" (Webster's dictionary), while discovery, implies the self initiated search for and acquisition of new knowledge. The last D stands for data mining. Data mining is a set of methods (or algorithms) for digging up patterns from data.

Hence KDD would broadly refer to the "process of finding knowledge in data [with] the 'high level' application of particular data mining methods." This implies that data mining working independently without some KDD framework is not able to produce useful knowledge.

Read the report about the original KDD workshop by Gregory Piatetsky-Shapiro:

Here's a good definition of what data mining is