基于自适应权重模糊C-均值聚类的瓦斯突出预测

Prediction on Gas Outburst of Fuzzy C-Mean Value Cluster base on Self Adaptive Weighting

传统的模糊C-均值聚类在处理煤与瓦斯突出的预测时,由于其对初始聚类中心的过度依赖而导致预测结果准确率的降低。为了准确预测煤与瓦斯突出,提出了一种基于自适应权重模糊C-均值聚类的煤与瓦斯突出预测方法。该方法将瓦斯浓度相关影响因素作为特征空间中的样本,利用高斯距离比例来表示权重,并动态计算每个样本对于类的权重,对特征空间中的样本进行聚类分析预测,降低了算法对初始聚类中心的依赖。

When the traditional fuzzy C-mean value cluster was applied to predict the coal and gas outburst,due to the over dependent on the initial cluster center caused a reduction in the accuracy of the predicted results,in order to accurately predict the coal and gas outburst,a prediction method of the coal and gas outburst was provided base on the fuzzy C-mean value cluster of the self adaptive weighting.The method applied the related influence factors of the gas content as the sample of the feature space.The Gaussian distance scale as the weighting was applied to dynamically calculate the cluster weight of each sample.The cluster analysis and prediction was conducted on the samples in the feature space and the algorithm dependent on the initial cluster center was reduced.