| 摘要 |
本发明提供一种样本属性的动态分布资料获取方法及系统，该方法包括：获取大规模样本中任一样本的样本属性；确定所获取的样本属性在样本属性伫列中的更新位置，在该更新位置上更新所获取的样本属性；保持样本属性伫列中的样本属性的有序排列，得到样本属性的动态分布资料；样本属性伫列中储存大规模样本中的样本属性，样本属性在该伫列中有序排列，该伫列长度为N，N小于大规模样本中的总样本数。本发明基于简单随机抽样原理，只需维持长度为N的样本属性伫列即可得到样本属性的动态分布资料，减小样本属性的动态分布资料获取的计算量。; ascertaining an updated position of the acquired sample attributes in a pre-sustained sample attribute sequence, updating the acquired sample attributes in the updated position; keeping sample properties in the sample attributes sequence in order, acquiring dynamic distribution data of sample attributes; the sample attribute sequence stores properties of samples of the large scale samples, sample properties in the sample properties sequence are arranged in order, and length of sample attributes are set up as N, which is smaller than the number of samples in the large scale samples. Based on simple random sampling principle, the present invention acquires dynamic distribution data of sample attributes as long as sample attribute sequence keeps length as N, therefore calculation of dynamic distribution data of sample attribute is reduced. |
