Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Comparison of two inductive learning methods: A case study in failed fuel identification

Conference ·
OSTI ID:5350098
 [1];  [2]
  1. Argonne National Lab., IL (United States)
  2. Michigan Univ., Ann Arbor, MI (United States). Dept. of Nuclear Engineering

Two inductive learning methods, the ID3 and Rg algorithms, are studied as a means for systematically and automatically constructing the knowledge base of expert systems. Both inductive learning methods are general-purpose and use information entropy as a discriminatory measure in order to group objects of a common class. ID3 constructs a knowledge base by building decision trees that discriminate objects of a data set as a function of their class. Rg constructs a knowledge base by grouping objects of the same class into patterns or clusters. The two inductive methods are applied to the construction of a knowledge base for failed fuel identification in the Experimental Breeder Reactor II. Through analysis of the knowledge bases generated, the ID3 and Rg algorithms are compared for their knowledge representation, data overfitting, feature space partition, feature selection, and search procedure.

Research Organization:
Argonne National Lab., IL (United States)
Sponsoring Organization:
DOE; USDOE, Washington, DC (United States)
DOE Contract Number:
W-31109-ENG-38
OSTI ID:
5350098
Report Number(s):
ANL/CP-74367; CONF-920538--16; ON: DE92011843
Country of Publication:
United States
Language:
English