摘要 |
<p>Apparatus for generating association rules for use in ensemble machines in data mining, comprising: universal set constructing means for constructing a universal set S of fundamental rules for constructed ensembles; rule canceling means for canceling rules of nodes higher than a node generating a relevant rule for all the rules that are selected to meet external input conditions, from the constructed universal set S; confidence calculating means for calculating confidences of the rules that belong to the universal set S from which the higher nodes have been canceled by the rule canceling means; arranging means for arranging the rules within the universal set S in order of the confidences calculated by the confidence calculating means; similarity comparing means for comparing similarity between the rules by using the rules of the universal set S that are arranged by the arranging means; and data analyzing means for analyzing data in order of the confidences based on the compared similarity.</p> |