基于支持向量机的核爆地震自动识别

针对核爆地震识别问题的特点,提出利用支持向量机(SVM)方法进行核爆地震的自动识别。该方法借助算法的内在能力来实现特征的选择变换,不必像传统方法那样将很大的精力用于特征空间的降维处理。同时,由于该方法建立在结构风险最小化准则上,而不是仅仅使经验风险最小,所以,它具有好的推广能力。实际数据处理结果表明,该方法在小样本情况下性能优于神经网络,可以很好地克服过学习问题。

Recognition of nuclear explosion and natural earthquake based on SVM

In this paper, Support Vector Machines (SVM) method is introduced to solve a two classes classification problem , on which seismic recordings of nuclear explosions must be distinguished from the recordings of natural earthquakes. Using SVM one needn't waste time on feature extraction but resort to the intrinsic feature extraction ability, which makes it more practical. The learning discipline of SVM is to minimize the structural risk instead of empirical risk, hence the better extensibility is guaranteed. The result of practical application indicates that the performance of SVM has superiority over ANN and can overcome the problem of "over fitting" excellently.