spectral clustering
1416
IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS, VOL. 24, NO. 9, SEmEMBER 1994
Instead of probing at the inputs and outputs of the first candidate, we can replace it with a circuit board that is known to be good. Then, we apply the same test to the primary inputs and compare the output vector after the replacement with the old one. If the two vectors are identical, then the original board was not faulty and diagnosis proceeds with the remaining candidates. If they are different and the new one is correct, then the original board was faulty and diagnosis can be terminated. Otherwise, the original board was faulty and there are still unknown faults. The diagnosis continues on the remaining candidates reordered according to the new symptom. The result of applying a different test to the device can also be used to reorder candidates. For example, we can move candidates connecting to corroborations under the new test to the end of the candidate list and move candidates connecting to violations to the front. The ordered candidate list is sufficient for selecting the best action among the same type but is not sufficient for selecting actions of different types. To select actions among different types a thorough analysis of probabilities and action costs is necessary[2, 41.
[18] W. B. Rouse,“A model of human decision making in a fault diagnosis task,” IEEE Trans. on Syst., Man, and Cybem., SMC-8, no. 5, pp. 357-361, May 1978.[19] W. B. Rouse,“Problem solving performance of first semester maintenance trainees in two fault diagnosis tasks,” Human Factors, vol. 21, no. 5, pp. 611-618, Oct. 1979.[20] W. B. Rouse,“Problem solving performance of maintenance trainees in a fault diagnosis task,” Human Factors, vol. 21, no. 2, pp. 195-203, Apr. 1979.[21] W. B. Rouse,“Models of natural intelligence in fault diagnosis tasks: Implications for training and aiding of maintenance personnel,” in Proceedings of the Joint Services Workshop on Artificial Intelligence in Maintenance, pp. 193-212, Jun. 1984.[22] W. B. Rouse, S. H. Rouse and S. J. Pellegrino,“A rule-based model of human problem solving performance in fault diagnosis tasks,” IEEE Trans. on Syst., Man, and Cybem., SMC-10, no. 7, pp. 366-376, Jul.
1980.[23] S. C. Shapiro, S. N. Srihari, M. R. Taie and J. Geller, VMES:“A network-based versatile maintenance expert system,” in Proceedings of the 1st International Conference on Applications of AI to Engineering Problems, pp. 925-936. Springer-Verlag, Apr. 1986.[24] M. R. Taie, J. Geller, S. N. Srihari and S. C. Shapiro,“Knowledge based modeling of circuit boards,” in Proceedings of 1987Annual Reliability and Maintainability Symposium, pp. 422427, Jan. 1987.[25] Z. Xiang,“Multi-level Model-based Diagnostic Reasoning,” Ph.D. Thesis, Department of Computer Science, University at Buffalo, Buffalo, NY 14260, Aug. 1988, Technical Report 88-17.[26] Z. Xiang and S. N. Srihari,“A strategy for diagnosis based on
empirical and model knowledge,” in Proceedings of the Sixth Intemational Workshop on Expert Systems and Their Applications, pp. 835-848, Avignon,
REFERENCES [l] B. Chandrasekaran and R. Milne,“Reasoning about structure, behavior
and function,” ACM SIGART Newsletter, vol. 93, Jul. 1985.[2] J. S. Chen,“A Probabilistic Theory of Model-Based Diagnosis,” Ph.D. Thesis, Department of Computer Science, University at Buffalo, Buffalo, NY, 1992, Technical Report 92-04.[3] J. S. Chen and S. N. Srihari,“Candidate ordering and elimination f in model-based fault diagnosis,” in Proceedings o the Eleventh International Joint Conference on ArtiJcial Intelligence, pp. 1363-1368, Morgan Kaufmann, 1989.[4] J. S. Chen and S. N. Srihari,“Action selection in interactive model-based diagnosis,” in Proceedings of the Eighth IEEE Conference on Artificial Intelligence Applications, pp. 67-73, 1992.[5] H. C. A. Dale,“Fault-finding in electronic equipment,” Ergonomics, vol. 1, pp. 356385, 1957.[6] R. Davis,“Diagnostic reasoning based on structure and behavior,” Artificial Intelligence, vol. 24, no. 3, pp. 347410, 1984.[7] R. Davis and W. Hamscher,“Model-based reasoning: Troubleshooting,” in H. Shrobe, editor, Exploring Artificial Intelligence: Survey Talksfrom the National Conferences on Artificial Intelligence, ch. 8, pp. 297-346. Morgan Kaufmann, San Mateo, CA, 1988.[SI R. Davis, H. Shrobe, W. Hamscher, K. Wieckert, M. Shirley and S. Polit,“Diagnosis based on description of structure and function,” in Proceedings of the Second National Conference on Artificial Intelligence,
France, Apr. 1986.
Finding a Subset of Representative Points in a Data Set D. Chaudhuri, C. A. Murthy, and B. B. Chaudhuri Abstract-This correspondence deals with the problem of finding the representative points from a data set 5 Rz.W o algorithms are stated. One of the algorithms can find the local best representative or seed points. The extension of these algorithms for three or more dimensions is also discussed. Experimental results on synthetic and real life data are provided which manifest the utility of these algorithms.
pp. 137-142. Morgan Kaufmann, 1982.[9] J. de Kleer,“An assumption-based TMS,” Artificial Intelligence, vol. 28, pp. 127-162, 1986.[lo] J. de Kleer and B. C. Williams,“Diagnosing multiple faults,” Artificial Intelligence, vol. 32, no. 1, pp. 97-130, 1987.[l 11 M. R. Genesereth,“The use of design description in automated diagnosis,” Artificial Intelligence, vol. 24, no. 3, pp. 41 1436, 1984.[12] R. M. Hunt and W. B. Rouse,”A fuzzy rule-based model of human problem solving,” IEEE Trans. on Syst., Man, and Cybem., SMC-14, no. 1, pp. 112-120, Jan./Feb. 1984.[I31 J. Rasmussen and A. Jensen,“Mental procedures in real-life tasks: A case study of electronic trouble shooting,” Ergonomics, vol. 17, no. 3, pp. 293-307, 1974.[14] J. A. Reggia, D. S. Nau and P. Y. Wang,“Diagnostic expert system based on a s
et covering model,” International Journal of Man-Machine Studies, vol. 19, pp. 437-460, 1983.[15] J. A. Reggia, D. S. Nau and P. Y. Wang,“A formal model of diagnostic inference. I. Problem formulation and decomposition,” Information Sciences, vol. 37, pp. 227-256, 1985.[16] R. Reiter,“A theory of diagnosis from first principles,” Artificial Intelligence, vol. 32, no. 1, pp. 57-95, 1987.[17] W. B. Rouse,“Human problem solving performance in a fault diagnosis task,” IEEE Trans. on Syst., Man, and Cybem., SMC-8, no. 4, pp. 258-271, Apr. 1978.
I. INTRODUCTION Consider a set of objects represented as a point data in Rz feature space. Given n data, we address the problem of selecting a small subset of no<< n data that faithfully represents the spatial organization of the original data. The solution to this problem can find 1] applications in data compression, data clustering[8],[ 4,pattern classification, as well as statistical parameter estimation[9],[ 141. For example, in clustering, many algorithms start with a few seed points, where each seed point represents the core of one cluster. The minimum distance classifier or the S-nearest neighbor classifier needs the seed points as the best patterns representing the classes. Quite often, the best representative point may be the mode of the pattern set. Density and mode estimation are two classical problems in statistics. In many situations, the problem of finding the best Manuscript received June 22, 1992; revised August 26, 1993. The authors are with the Electronics and Communication Sciences Unit, Indian Statistical Institute, 203 Barrackpore Trunk Road, Calcutta 700 035, India. IEEE Log Number 9401467.
0018-9472/94$04.00 0 1994 IEEE

each data object is in exactly one subset ? ... – A cluster is a set of points such that ...“representative” point of a cluster 4 center-...
Analysis is the process of finding a ? subset D’ of D [D: data set to be analyzed...Sampling (determines a representative subset of the...
FewSparse Modeling for Finding Representative Objects_文学研究_人文社科_专业资料...few representatives for a dataset, i.e., a subset of data points that ...
Finding groups of objects such that the objects ...“representative” point of a cluster 4 center-... which is a subset of the points of the ...
each data object is in exactly one subset ? ...A cluster is a set of points such that any ...“representative” point of a cluster 4 center-...
each data object is in exactly one subset ? ... – A cluster is a set of points such that ...“representative” point of a cluster 4 center-...
each data object is in exactly one subset ? ...A cluster is a set of points such that any ...“representative” point of a cluster 4 center-...
? Finding groups of objects such that the ...“representative” point of a cluster 4 center-...(a) Green points denote the data set in a ...
? Finding similarities between data according to ...K-Medoids Clustering: Find representative objects (...points in the test set and the closest centroids...
‘gender’ and a representative population of ...data points are summarized in a single figure ?...Selection may involve the chossing a subset of ...
热门文档