Publisher's Synopsis
Book Description: We have the data! Too much data! Decision-making requires that data be filtered and refined to provide information. Adding context to the content produces actionable knowledge. Unfortunately, current techniques strip away the uncertainty associated with the raw data. This book provides a decision-centered approach for coping with uncertainty that combines what people do best with what computers do best. Algorithms "plug into"the knowledge base from a single import/export interface, facilitating multi-strategy reasoning. Triage filters the data, extraction of hedge words capture uncertainty, an executable knowledge base provides content in context, data fusion propagates uncertainty, data analytics discover patterns, and plan optimization tools move the decision-maker from "what's going on" to "what to do". Displays present actionable knowledge with associated uncertainties explicitly shown. Fifteen applications are shown ranging from longevity prediction, to a retail problem solver, to intelligence community applications, to starship cybernetics. We wrote the book to provide the practitioner with compelling ideas for orchestrating artificial intelligence, statistical, and mathematical algorithms to produce fully integrated decision support systems. Novel techniques of particular interest are: a knowledge representation that provides a unifying framework for multi-strategy reasoning and simulation, a robust treatment of uncertainty, monitor-assess-plan-execute decision loops for routine and quick-reaction decisions, eight techniques for automated discovery of unknown unknowns, level 4 (process refinement) data fusion, and a self-aware knowledge base that "knows what it knows".
