On the Classification of Patterns by the Karhunen-Loeve Orthogonal System Without Supervisor
Proceedings of the Second International Joint Conference on Artificial Intelligence. Pages 424-427.
Abstract
In this paper we show how the pattern samples generated from the unknown asymmetric finite mixture distribution are dichotomized by the Karhunen-Loeve orthogonal system in the optimal wa y. We introdu ce the concept of the difference of features of two categories and prove that decreasing the error proba b ilit y of dichotom y is equal to increasing the difference of features of two categ o ries. And it can be shown that the maximum difference of features obtained by the KL system is equivalent t o that which is obtained by the Ba y es soluti o n. Moreover it can be pro ved that there exists a dichotom y which converges to the Bayes s o lution by increasing the number of samples.
@inproceedings{ijcai1971p44,
title = {On the Classification of Patterns by the Karhunen-Loeve Orthogonal System Without Supervisor},
author = {Tomita, Shingo and Noguchi, Shoichi},
booktitle = {Proceedings of the Second International Joint Conference on
Artificial Intelligence, {IJCAI-71}},
publisher = {International Joint Conferences on Artificial Intelligence},
editor = {D.C. Cooper},
pages = {424--427},
year = {1971},
url = {https://ijcai.org/proceedings/1971/44},
}
TY - CPAPER TI - On the Classification of Patterns by the Karhunen-Loeve Orthogonal System Without Supervisor AU - Tomita, Shingo AU - Noguchi, Shoichi T2 - Proceedings of the Second International Joint Conference on Artificial Intelligence, IJCAI-71 PB - International Joint Conferences on Artificial Intelligence PY - 1971 SP - 424 EP - 427 UR - https://ijcai.org/proceedings/1971/44 ER -Copied
Machine-readable versions of this paper, for tools and AI agents:
