CAJ | 학술논문

针对小样本、多类煤矿顶板的状态检测问题，提出了基于支持向量机集成的算法。采用深度优先搜索对支持向量机集成参数进行优化，并结合煤矿项板敲击声信号的人耳听觉功率谱特征，对煤矿顶板中浮石、剥层等现象进行分类识别。实验表明，该算法能够对多类顶板状态进行有效分类，且识别率较高，能够作为安全开采的保障手段。
침대소양본、다류매광정판적상태검측문제，제출료기우지지향량궤집성적산법。채용심도우선수색대지지향량궤집성삼수진행우화，병결합매광항판고격성신호적인이은각공솔보특정，대매광정판중부석、박층등현상진행분류식별。실험표명，해산법능구대다류정판상태진행유효분류，차식별솔교고，능구작위안전개채적보장수단。
In this paper, a method based on support vector machine ensemble about the detection of Many types of coal mine roof is proposed. In optimization process, the depth-first search is used. The classification and recognition of security risk is according to power spectral characteristics of human ear of mine roof percussion signal. The experiments show that the algorithm can classify the security risk of coal mine roof effectively.