【摘要】 Log Gaussian Cox过程模型(简称LGCP模型)常用来描述关于空间变化的随机过程,但是它不能很好地拟合强度取对数后是非高斯过程情况下的数据。因此,通过将它的强度取对数后看成是一个混合高斯过程来改进LGCP模型,并研究改进后模型的性质。采用极大似然估计法和MCMC方法来估计模型参数,以及用AIC准则作模型选择。最后通过实例验证,结果显示改进后的模型能够有效地拟合数据。更多还原

【Abstract】 The Log Gaussian Cox Process Model( referred to as the LGCP model) is often used to describe spatial stochastic processes,but it cannot fit data well where its logarithmic intensity is the non-Gaussian process. Thus,the LGCP model is improved by considering the logarithmic intensity as a Gaussian mixture process,and the property of the improved model is derived. The model parameters are estimated by the maximum likelihood estimation and the MCMC method,and model selection is done through the AIC criterion. Finally,the result of the example shows that the improved model can fit the data more effectively.更多还原