自适应免疫算法的SVME用于水下目标识别

针对用支持向量机集成提高水下目标识别正确率会使识别系统更加复杂的问题,提出了一种以自适应免疫算法(AIA)的支持向量机选择性集成(SVME)算法(即AIA-SVME算法)进行分类器优化选择,对实测水下目标声信号进行分类识别.与分类器全部集成的识别实验对比证明,该算法在选择9%的分类器后仍可以达到分类器全部集成的识别效果,不仅保证了识别精度,还使得识别系统大幅度精简,节省在线识别的时间.该研究对于水下目标分类决策优化集成的新方法探索具有重要理论价值和实际意义.

Underwater target recognition based on AIA-SVME algorithm

A self-adaptive immune algorithm (AIA) based supporting-vector-machine election ensemble (SVME) algorithm, i.e. the AIA-SVME algorithm, is proposed for optimal selection of classifiers. And it is applied to classifying underwater targets. Comparing the classification by the proposed algorithm with that by assembling all the classifiers, we find that the same classification effect is achieved by only choosing 9% of the classifiers. So the AIA-SVME algorithm can not only guarantee the classification accuracy, but also simplify the recognition system greatly.