@article{oai:bunkyo.repo.nii.ac.jp:00003280,
 author = {鈴木, 昇一},
 journal = {情報研究, Information and Communication Studies},
 month = {1981-01-01, 2012-01-17},
 note = {Let a recognition system secure an internal model in memory corresponding to an input pattern. If the internal model can be constructed in such a way that the system extracts from the model the same features that input pattern has, the model is called a reduced structural model. In this paper, giving a good explanation of definitions and the resulting necessity thus far we calculate features and the binary features extracted from a structural pattern which possesses a more general form structurally than the internal model and determine the upper limits and the lower limits of the features.
We also explain a meaning and its application of a structural pattern of the same type called the most-approximate structural pattern that satisfies the condition that a best degree of approximation of the norm-normalized input pattern may be obtained. Lastly we seek an output pattern from a Perceptron-like spatial circuit to which the structural pattern is applied and it is shown that the output pattern has the same form as the input structural pattern.},
 pages = {13--31},
 title = {パターン情報処理における構造化パターン,最良近似構造化パターンと簡約構造モデル},
 volume = {2},
 year = {}
}