作者: Jitender Deogun , Liying Jiang
DOI: 10.1007/11548706_11
关键词:
摘要: An interesting application of association mining in the context temporal databases is that prediction. Prediction to use antecedent a rule predict consequent rule. But not all rules may be suitable for In this paper, we investigate properties prediction, and develop an approach called prediction — set are useful discovers have three properties. First, there must time lag between Second, minimum condition implies consequent. Third, relatively stable confidence with respect frame determined by domain. We algorithm discovering rules. The efficiency effectiveness our validated experiments on both synthetic real-life databases, show efficiently proper