很好的一本书


These workshop notes, written in 2008, cover statistical methods available in public domain software.
The workshop uses the statistical package 'R' and is based on 'spatstat', an add-on library for 'R' for the analysis of spatial data.
Topics covered include:

• statistical formulation and methodological issues
• data input and handling
• R concepts such as classes and methods
• nonparametric intensity estimates
• goodness-of-fit testing for Complete Spatial Randomness
• maximum likelihood inference for Poisson processes
• model validation for Poisson processes
• distance methods and summary functions such as Ripley’s K function
• non-Poisson point process models
• simulation techniques
• fitting models using summary statistics
• Gibbs point process models
• fitting, simulating and validating Gibbs models
• multitype and marked point patterns
• exploratory analysis of marked point patterns
• multitype Poisson process models and maximum likelihood inference
• multitype Gibbs process models and maximum pseudolikelihood
• line segment data.
  

This workshop requires 'R' version 2.6.0 or later, and 'spatstat' version 1.14-5 or later.