基于Rough Sets的传感器异常数据处理

在各种传感器的应用中,经常要对传感器的测量数据进行处理,以保证测量结果的可靠性．为了利用粗糙集理论处理不确定数据的优点,根据粗糙集理论的思想,先由已知测量数据提取出决策表,再进行补全、离散化等预处理,最后进行属性约简并提取出分类规则,对测量数据进行分类,剔除测量数据中的异常数据．实验结果显示该异常数据发现方法比常用的异常数据处理方法更为客观、精确和可靠．

Abnormal Data Processing of Sensor Based on Rough Sets

In the applications of various sensors,the measurement data are usually to be processed in order to ensure the reliability and the accuracy of the measurement result.To make use of the advantage of Rough sets in uncertain data processing and according to the Rough sets theory idea,some pretreatment steps,such as obtaining decision table from the known measurement data,completing decision table and diseretization,should be carried out.And then after reducing the attributes and extracting the classification rule,the measurement data are classified and the abnormal data are eliminated.The experimental result shows that this method is more objective, accurate and reliable than usual method of finding abnormal data.