A dynamic parallel forecasting model is proposed, which is based on the problem of current forecasting models and their combined model. According to the process of the model, the fuzzy C-means clustering algorithm is ...A dynamic parallel forecasting model is proposed, which is based on the problem of current forecasting models and their combined model. According to the process of the model, the fuzzy C-means clustering algorithm is improved in outliers operation and distance in the clusters and among the clusters. Firstly, the input data sets are optimized and their coherence is ensured, the region scale algorithm is modified and non-isometric multi scale region fuzzy time series model is built. At the same time, the particle swarm optimization algorithm about the particle speed, location and inertia weight value is improved, this method is used to optimize the parameters of support vector machine, construct the combined forecast model, build the dynamic parallel forecast model, and calculate the dynamic weight values and regard the product of the weight value and forecast value to be the final forecast values. At last, the example shows the improved forecast model is effective and accurate.展开更多
为更加快速准确地从微震时序数据中提取微震事件,提高异常事件的捕捉效率,提出一种基于多尺度融合卷积和空洞卷积的自动编码器(multi-scale fusion convolution and dilated convolutions auto encoder,MDCAE)与融合波动率和限制窗口的...为更加快速准确地从微震时序数据中提取微震事件,提高异常事件的捕捉效率,提出一种基于多尺度融合卷积和空洞卷积的自动编码器(multi-scale fusion convolution and dilated convolutions auto encoder,MDCAE)与融合波动率和限制窗口的动态时间扭曲(constraints dynamic time warping for fusing volatility,CDTW-Vol)方法。提出MDCAE的特征提取方法,将波形信号转变为低维特征信号,引入微震波形的波动率的概念,通过改进后的DTW算法对特征信号进行相似性度量,得到的相似性矩阵进行k-medoids聚类,得到聚类结果。应用某矿区501工作面和802工作面微震监测数据集进行实验,验证所提方法的准确性和泛化性,经实验得出所提聚类方法轮廓系数89%,兰德系数90%,相比普通的k-medoids聚类算法聚类精度上升57%,为捕捉微震系统的异常事件提供了一种新方法。展开更多
It is known that centers, widths, and weights are three mainly considered factors in constructing a radial basis function（RBF） network.This paper aims at constructing a compact RBF network with two main steps.In the...It is known that centers, widths, and weights are three mainly considered factors in constructing a radial basis function（RBF） network.This paper aims at constructing a compact RBF network with two main steps.In the first step, the coarse clusters computed from triangle inequalities are refined to obtain the locations of centers by the defined maximum degree spanning tree（MDST）.Meanwhile the coarse widths are obtained.In the second step, a learning algorithm referred to as anisotropic gradient descent method is presented to further refine the above coarse widths.Experiments of the proposed algorithm show its great performance in times series prediction and classification.展开更多
基金supported by the National Defense Preliminary Research Program of China(A157167)the National Defense Fundamental of China(9140A19030314JB35275)
文摘A dynamic parallel forecasting model is proposed, which is based on the problem of current forecasting models and their combined model. According to the process of the model, the fuzzy C-means clustering algorithm is improved in outliers operation and distance in the clusters and among the clusters. Firstly, the input data sets are optimized and their coherence is ensured, the region scale algorithm is modified and non-isometric multi scale region fuzzy time series model is built. At the same time, the particle swarm optimization algorithm about the particle speed, location and inertia weight value is improved, this method is used to optimize the parameters of support vector machine, construct the combined forecast model, build the dynamic parallel forecast model, and calculate the dynamic weight values and regard the product of the weight value and forecast value to be the final forecast values. At last, the example shows the improved forecast model is effective and accurate.
文摘为更加快速准确地从微震时序数据中提取微震事件,提高异常事件的捕捉效率,提出一种基于多尺度融合卷积和空洞卷积的自动编码器(multi-scale fusion convolution and dilated convolutions auto encoder,MDCAE)与融合波动率和限制窗口的动态时间扭曲(constraints dynamic time warping for fusing volatility,CDTW-Vol)方法。提出MDCAE的特征提取方法,将波形信号转变为低维特征信号,引入微震波形的波动率的概念,通过改进后的DTW算法对特征信号进行相似性度量,得到的相似性矩阵进行k-medoids聚类,得到聚类结果。应用某矿区501工作面和802工作面微震监测数据集进行实验,验证所提方法的准确性和泛化性,经实验得出所提聚类方法轮廓系数89%,兰德系数90%,相比普通的k-medoids聚类算法聚类精度上升57%,为捕捉微震系统的异常事件提供了一种新方法。
基金supported by Key Program of National Natural Science Foundation of China (U0635001)China Postdoctoral Science Foundation (20060390728)the Natural Science Fund of Guangdong Province, China (07006490)
文摘It is known that centers, widths, and weights are three mainly considered factors in constructing a radial basis function（RBF） network.This paper aims at constructing a compact RBF network with two main steps.In the first step, the coarse clusters computed from triangle inequalities are refined to obtain the locations of centers by the defined maximum degree spanning tree（MDST）.Meanwhile the coarse widths are obtained.In the second step, a learning algorithm referred to as anisotropic gradient descent method is presented to further refine the above coarse widths.Experiments of the proposed algorithm show its great performance in times series prediction and classification.
