Inductive Learning of Categories from Examples Using Minimum Cost Representations
Abstract
The problem of learning categories from a sequence of examples is considered in terms of maintaining a minimum-cost representation for the set of categories. Categories here are subsets of a set of natural numbers. Computational aspects of the following problems are addressed: (1) how are the category representations updated as an incremental change is made to one category, (2) how can a minimum cost representation for a new category be obtained in terms of existing ones, and (3) how does the enlargement of the known universe of objects affect representations of known categories. We then discuss extensions of our methodology to domains which include structure in the categorical descriptions.
Cite
Text
Tanimoto. "Inductive Learning of Categories from Examples Using Minimum Cost Representations." International Joint Conference on Artificial Intelligence, 1979.Markdown
[Tanimoto. "Inductive Learning of Categories from Examples Using Minimum Cost Representations." International Joint Conference on Artificial Intelligence, 1979.](https://mlanthology.org/ijcai/1979/tanimoto1979ijcai-inductive/)BibTeX
@inproceedings{tanimoto1979ijcai-inductive,
title = {{Inductive Learning of Categories from Examples Using Minimum Cost Representations}},
author = {Tanimoto, Steven L.},
booktitle = {International Joint Conference on Artificial Intelligence},
year = {1979},
pages = {871-873},
url = {https://mlanthology.org/ijcai/1979/tanimoto1979ijcai-inductive/}
}