中文摘要：

针对影响模式识别特征向量处理和分析的样本冗余数据问题，提出一种基于ISODATA算法的样本冗余数据辨识和剔除方法．该方法能够根据剔除精度检测出样本中的冗余数据并剔除．对实测数据应用该方法进行辨识，结果表明，该方法能够调整数据剔除精度，从原始特征中挑选出最有代表性、分类性能最好的特征．

英文摘要：

To eliminate the influence of redundant data in feature vector of pattern recognition processing and analysis,the paper presents a method for identification and elimination of redundant data by design of a clustering system. The method according to precise degree detects and removes redundant data. The results show that the method can be combined with the expertise to select the most representative from the original features, the classification performance of the best features of the measured data using this method.