针对当前的主要预警方法的问题,提出基于主动学习的支持向量数据描述预警技术,并给出了该技术的实现过程。该技术利用支持向量数据描述解决了缺少预警警兆数据的问题,利用主动学习解决了正常样本和待测样本数据量大的问题,以某地区的宏观经济运行预警系统为例,证明了该技术的有效性。
In order to solve the problem of inapplicability of current early warning methods, a support vector data description early warning technique based on active learning is proposed. And implementation of the technique is detailedly presented. In the technique, support vector data description is used to choose "ill-represented" samples and active learning was used to cope with the the large size of observations. An early warning experiment is conducted on a macro-economic dataset to verify the proposed technique.