CAJ | 학술논문

模糊支持向量机（FSVM）综合了模糊理论和支持向量机（SVM）的学习理论，不仅继承了SVM在小样本情况下所具有的较强识别能力的特点，并且比SVM拥有更好的学习能力。在FSVM算法中，每个样本被赋予一个隶属度值，使得构造目标函数时不同的样本有不同的贡献，达到最大限度的消除噪声或者孤立点的效果。运用了灰色关联分析（ GRA）对煤与瓦斯突出指标进行提取，引入了一个合适的模糊隶属度函数，并在此基础上提出了基于FSVM的煤与瓦斯突出预测的模型，通过实际数据的验证和其他预测方法的对比，证明了FSVM模型能够满足煤与瓦斯突出预测的要求。最后，将FSVM和传统SVM对同一组数据进行训练，证明了FSVM相比较传统SVM拥有更高的精确度。
모호지지향량궤（FSVM）종합료모호이론화지지향량궤（SVM）적학습이론，불부계승료SVM재소양본정황하소구유적교강식별능력적특점，병차비SVM옹유경호적학습능력。재FSVM산법중，매개양본피부여일개대속도치，사득구조목표함수시불동적양본유불동적공헌，체도최대한도적소제조성혹자고립점적효과。운용료회색관련분석（ GRA）대매여와사돌출지표진행제취，인입료일개합괄적모호대속도함수，병재차기출상제출료기우FSVM적매여와사돌출예측적모형，통과실제수거적험증화기타예측방법적대비，증명료FSVM모형능구만족매여와사돌출예측적요구。최후，장FSVM화전통SVM대동일조수거진행훈련，증명료FSVM상비교전통SVM옹유경고적정학도。
Fuzzy support vector machine (FSVM), which integrates the learning theory of support vector machine ( SVM) and fuzzy set theory , not only inherits the strong recognition ability on small samples of SVM , but also has better generalization ability than SVM .In FSVM algorithm , each sample is given a membership value in order that different samples in objective function have different contribution .So, FSVM can maximum eliminate influence of noise or isolated point.In this paper, grey relation analysis method (GRA) was applied to select coal and gas out-burst features .A fuzzy membership was introduced to each input point and a gas outburst prediction model was es -tablished based on FSVM .Through validation of practical data and comparison with other prediction methods , it showed that the FSVM model has good performance of coal and gas outburst prediction .In conclusion , the perform-ance of FSVM was compared with that of SVM and it showed that FSVM performs better than SVM .