可拓模式识别算法中经典域的确定方法

The Method of Determining Classical Domain with the Extension Pattern Recognition Algorithm

摘要:

可拓模式识别算法是根据各个关联度的对比来识别最终的模式,其中经典域区间的确定有着重要的作用。传统的经典域确定方法是基于数理统计学的,不适用于小样本数据,数据不完全以及数据分布未知等复杂情况。引入 Bootstrap算法,并和传统的方法对比,更精确地确定了经典域区间。

Abstract:

The extension pattern recognition algorithm is to identify the final pattern by comparing the correlation degree. Inthis process, the determination of classical domain is important. The traditional methods of determining the classical do.main were based on mathematical statistics, and were not suitable for the small sample data, incomplete data and unknowndistribution and other complex. By introducing the new algorithm, compared with the traditional methods, the new algo.rithm determined a more precise classical domain interval.