delta learning system for using specialistadvice on revising fehlerhierarchien in diagnostic expert systems

摘要

A computer system takes as input an initial fault hierarchy KB0 and a set of annotated session transcripts C = ä&lang& pi j, rj&rang&ü and is given a specified set of revision operators T = äiü where each i ELEMENT T maps a fault hierarchy KB to a slightly different hierarchy i(KB). This revised fault hierarchy i(KB) is called a neighbor of KB, and the set of all neighbors of KB, N(KB)=äi(KB) ü, is KB'S neighborhood. The computer system uses T to hill climb from the initial fault hierarchy KB0, through successive hierarchies, KB1... KBm, with successively higher empirical accuracies over C. The final hierarchy, KBm, is a local optimum in the space defined by the revision operators. At each stage, to go from a fault hierarchy KBi to its neighbor KBi+1, the present invention must evaluate KBi's accuracy over C, as well as the accuracy of each KB' ELEMENT N(KBi). If any KB' is found to be more accurate than KBi, this KB hierarchy becomes the new standard, labeled KBi+1, and the theory revision process iterates, seeking a neighbor of this KBi+1, that is more accurate than KBi+1 over the set of examples C, and so forth. Otherwise, if no KB' is more accurate than KBi, the hill climbing process will return this KBi, and terminate. IMAGE