中文摘要：

对基因表达数据进行分类时,超限学习机（ELM）算法具有学习效率高、泛化能力强、分类精度高的优点.为了解决超限学习机算法受输入权值矩阵和隐含层偏差随机初始化的影响,本文利用自适应遗传算法（AGA）具有良好的全局搜索效果对超限学习机的输入权值矩阵和隐含层偏差进行优化,提出了基于自适应遗传算法优化超限学习机（AGA-ELM）的分类算法.通过实验表明,该算法与已有的ELM、GA-ELM以及SVM算法相比,分类精度更高,可用于基因数据分类.

英文摘要：

Extreme learning machine algorithm （ELM） has high learning efficiency, high generalization capability and high classification accuracy for gene expression data classification. In order to avoid the side- effect of the random input layer weights and the hidden layer bias, an adaptive genetic algorithm（AGA） was integrated into the ELM algorithm to optimize the input layer weight matrix and the hidden layer bias. The new algorithm is called AGA-ELM. The experiment shows that the gene expression data classification results of AGA-ELM are higher than the algorithms such as ELM, GA-ELM and SVM.