节点文献
混合Weibull分布参数估计的EM算法
EM Algorithm for Mixture Weibull Distribute Density Estimation
【摘要】 混合Weibull分布在工业、经济、管理等领域中有着广泛的应用背景.但混合Weibull模型相当复杂,其参数估计相当困难,对于这一问题,目前还没有一种行之有效的解决方法.本文推导了求解有限混合Weibull模型参数估计的EM算法,并设计了一种有效选取参数初始值的方法.大量数值试验表明:文中给出的算法及初始化方案能有效地解决有限混合Weibull分布的参数估计问题.
【Abstract】 The mixture Weibull distribution can be widely applied in many areas such as industry, economy, management, etc. However, due to the complexity of mixture Weibull model, the problem of parameters estimation for mixture weibull has not been completely resolved yet . The EM algorithm (expectation-maximization algorithm) for mixture Weibull distribution is derived. The efficient initialization methods for these iterative formulas are presented. Experiental results of numerical simulation show that the more accurate results can be obtained by proposed initial methods.
【Key words】 mixture Weibull distribution; Maximum likelihood estimation (MLE); EM algorithm; complete data; incomplete data;
- 【文献出处】 华北工学院学报 ,Journal of North China Institute of Technology , 编辑部邮箱 ,2003年05期
- 【分类号】O211.67
- 【被引频次】20
- 【下载频次】462