Data vs. knowledge mining: a crossing of theories

Agent associates are knowledge-based systems that provide advice and guidance in (self-referentially) developing the knowledge base of an expert system. Practically speaking, agent associates extract critical knowledge from the user towards rule capture. The task of the knowledge engineer is thus reduced to that of random programming. In summary, this paper views knowledge mining as an extension of data mining. Whereas a loss of information is inherent to all stochastic data mining operations, a gain of random knowledge is inherent to all knowledge mining operations. Results have implication to directed data and knowledge mining. Results also shed new light on domain-specific approaches to cracking the knowledge acquisition bottleneck in knowledge-based systems