基于软件度量的数据挖掘结果的评估模型

The Evaluation Model of Data Mining Results Based on Software Measurement

对数据挖掘所产生的模式进行解释和评估是数据挖掘过程中的一个优化处理步骤,虽然可以通过设置模式的属性阈值消除大量无关模式,但当面对一个大的数据库时,数据挖掘的最终结果依然很大。从主观和客观两个方面分析模式评估的标准,提出了数据挖掘结果评估的度量构造,详细说明了各个度量的计算方法,对挖掘结果进行了综合处理。通过此模型可以对数据挖掘结果的可用性和可靠性进行客观的评价。

In data mining, the interpretation and evaluation on the patterns is a step of optimized processing. Although the massive irrelevant patterns can be eliminated through setting attribute threshold value, the results of data mining axe still large when dealing with a large database. This paper analyzes the standard of evaluating pattern from the aspects of subjective and objective, buihs the metric construction to the results of data mining, gives a detailed description for the calculation of each measure, and processes the results of data mining comprehensively. This model can be adopted in the objective evaluation on the availability and reliability of the data mining results.