矿井主运输系统火灾预测的RS-SVM模型

RS-SVM Model for Fire Prediction of Mine Transportation System

中文关键词:  火灾识别  粗糙集  支持向量机  贝叶斯算法  RBF-NN算法

英文关键词:fire identification  rough sets  support vector machines(SVM)

中文摘要:

      为了更高效、准确地预测矿井主运输传送带火灾的发生，提出了一种基于粗糙集-支持向量机RS-SVM的煤矿火灾预测算法。利用RS理论对8个变量映射为粗集知识系统进行离散化处理以及属性约简，去除冗余信息，排除对于实验不必要的干扰，获得知识系统规则集；通过训练确定RS-SVM模型，再回判来验证此模型的准则性，最后对RS-SVM、贝叶斯、RBF-NN三种预测算法进行样本的预测分析，结果表明RS-SVM算法与其他两种算法相比有着明显的优势，在少样本时的预测准确性更高、速度快、抗扰性好、非线性能力强，现场实用性强，使用范围广，对于火灾的预测具有重要意义。

英文摘要:

      In order to more efficiently and accurately predict the main transport of mine fire occurred with the transfer,a coal mine fire prediction algorithm based on rough set and support vector machine (RS-SVM) is put forward.Eight variables were mapped into rough set knowledge system for discretization and attribute reduction using RS theory to remove redundant information and exclude the unnecessary interference to experiments,obtain knowledge system rule set.The RS-SVM model is established through training,and the criterion of the model is verified by the return judgment.At the end,the sample prediction results of the three prediction algorithms of RS-SVM,Bayes and RBF-NN show that the RS-SVM algorithm has obvious advantages compared with the other two algorithms.It has higher prediction accuracy,faster speed,better anti disturbance,strong nonlinear ability,strong field practice and wide application range in small sample,and has important significance for the prediction of fire.