基于模式识别理论基坑监测数据周期性误差分析

摘要

要识别测量数据的周期性误差，传统解决方法是利用数理统计中的统计检验方法。论文结合模式识别理论和方法，以贝叶斯风险函数为基础，分析基坑监测数据的周期性误差。对比统计检验方法，模式识别方法的判识错误率最小，在识别测量数据是否有周期性误差的同时得到判识结果的可靠性。%To identify periodic error of measured data ,the traditional solution is to use statistical testing method of mathematical statistics .Based on pattern recognition theory and the bayesian risk function ,analysis of foundation pit mo-nitoring data .Contrast statistic test methods or pattern recognition method of minimum error rate on the identification of the measurement data if there is a periodic error to get to know the reliability of the results .