基于Fisher和核Fisher组合的磁暴识别方法

摘要

为从地磁水平分量序列中自动识别出磁暴,提出了一种基于Fisher和核Fisher组合的磁暴识别方法.使用北京十三陵地磁台的数据进行测试,在数据样本较少的情况下,该方法对于磁暴与非磁暴的平均识别率达到了85％以上,并随着样本数据的增多,识别率还将有所提高.%To recognize the magnetic storms automatically from the horizontal component of geomagnetism,a combined method of magnetic storms recognition based on Fisher and kernel Fisher discriminant analysis is presented.The data used in the study are derived from Ming tombs of Beijing geomagnetic station.In the case of the little number of the data samples,the average rate of this combined method for magnetic storms recognition is above 85 ％,and the recognition rate will be improved by the increasing amount of data samples.