Probabilistic Reasoning in Intelligent Systems

Chapter 8

LEARNING STRUCTURE FROM DATA

Publisher Summary

This chapter discusses the problems of constructing a network automatically from direct empirical observations, bypassing the human link in the process known as knowledge acquisition. The chapter discusses some of challenges posed by the problem. It chapter discusses the causal structures, particularly causal trees and poly trees, where the computational role of causality as a modularizer of knowledge achieves its fullest realization. The chapter focuses on the nature of causation, reiterating the thesis that causation is not a property of nature, but rather a mental construct devised for the efficient organization of knowledge. It presents some techniques for discovering causal structures ...