主元余像集主成分分析在蛋白质质谱数据中的应用

癌蛋白质谱数据中包含了大量未知的内部结构和变量。针对癌蛋白质谱数据这些特点，在总结主元余像集主成分分析（二次主成分分析）应用的基础上，提出了用t-验证方法进行特征子集选取，然后用主元余像集主成分分析提取特征，以线性判别分析进行分类的新方法。通过对典型癌蛋白质谱数据的分类实验，证明该方法不但识别率高，而且需要选取的特征子集小，分类速度快，提高了方法的准确性与分类速度。

Application of 2nd PCA on protein mass spectrometry data

Mass spectrometry data contain a large number of unknown internal structures and variables. This paper analyzes mass spectrometry data based on 2nd - PCA. T- test method is used for feature subset selection, 2nd - PCA is used for feature extraction further. And then, IDA is employed for the classification of mass Spectrometry data. With a classical set, we can not only prove the performance is superiority, but also prove the feature subset is smaller than others, the rapidity for classification is also quick. We can see the betterment of correct rate and rapidity clearly.