経験的予測のモデル化と歯車の振動推定に対する適用

概要

論文の詳細を見る
A hypothesis is assumed in that the empirical estimation reached by mechanical engineers in designing, trouble shooting and so on is worked out not by "rules" stored in the knowledge base which is deduced from their experiences, but by simple inter-or extrapolation of data concerning the objective matter in the memory of their experience, which are not conceptualized in the form of "rules". In this report a model for this is proposed. The procedure of prediction according to this model is as follows : 1. Specification of factors which influence the objective matter, and classification of concepts of nonnumeric factors among them in a tree structure. 2. Construction of the data base according to this conception structure. 3. Investigation of the degree of resemblance in objective matter between the investigating example and each example in the data base. 4. Decision of the estimating value by enveloping extrapolation of distributing points on the "objective matter vs degree of resemblance" chart. Predictions of gear vibration according to this model have proved that the accuracy of the estimation is quite good. When this model is incorporated in an expert system, it is no longer necessary for knowledge engineers to form "rules" by consulting experts. The job of knowledge engineers is to construct the attributes of the data base, that is, to construct the tree structure of the conception of factors which influence the objective matter.
一般社団法人日本機械学会の論文
1990-08-25

経験的予測のモデル化と歯車の振動推定に対する適用

スポンサーリンク

概要

著者

関連論文

スポンサーリンク