Abstract: A DNA microarray classification method based on Gaussian process is proposed.The method is a Bayesian classification algorithm that can give classification probability.The historical classification results can be easily added to the model to improve the model performance.In order to cope with the dimension increase，a dimensionality reduction method based on the principal samples is proposed.The method’s feasibility and effectiveness are proved in the comparison with several important classification methods.

Key words: microarray, Gaussian process, classification, statistical learning, Bayesian method

摘要： 基于高斯过程对分类过程进行建模，给出了一种基于高斯过程的DNA微阵列分类算法。作为一种贝叶斯分类方法，该方法能够给出分类的概率，并能将过往的正确诊断信息，纳入到分类模型中，实现分类模型的不断优化。该方法能够基于主样本进行训练空间的维度消减，较好地解决了由于样本的加入造成的维度不断增加的问题。通过和几种常用分类算法的实验对比分析，证明了该方法具有较高的分类准确性。

关键词: 微阵列, 高斯过程, 分类, 统计学习, 贝叶斯方法