做模式分类时（例如指纹识别，人脸识别），一个需要处理的难题是维数非常大，人脸往往是百万维的，目前计算机的能力还不足以快速地计算这么高维的数据。pca是一中降维的方法，用它可以把高维数据映射到一个维数较低的空间上考虑。-To do pattern classification (for example, fingerprint identification, face recognition), a need to deal with the problem is very large dimension, often millions of people face-dimensional, the current capacity of the computer fast enough to calculate such a high-dimensional data. PCA is a dimensionality reduction of the method, it can be high dimensional data is mapped to a lower dimension space to consider.