摘要 |
PROBLEM TO BE SOLVED: To provide a clustering method for grouping the production data of similar parameter operation groups by grouping data points satisfying a certain range having process operation similarity. SOLUTION: When a certain number of data points are given, a function block 41 forms a relational matrix. At a function block 42, a non-marked row having a single '1' is selected as the core of a cluster. At a function block 43, a non-marked column having the maximum number of non-marked columns is selected as the further core of the cluster. Next, a judge block 44 investigates the presence/absence of non-marked column. When all the columns are marked in this case, a function block 45 deletes all the columns, namely, the cores of the cluster corresponding to the selected row. At a function block 46, all the data points corresponding to '1' in the existent column of the selected row are added to these cores as cluster members for every selected row and next, the columns related to the selected cluster members of that row are deleted. |