This option fits various Non-Homogeneous Poisson Process (NHPP) models, described in detail below. Depending on the choices made in the launch window, the possible options are:
Once a model is fit, a Model List report appears. This report provides various statistical measures that describe the fit of the model. As additional models are fit, they are added to the Model List, which provides a convenient summary for model comparison. The models are sorted in ascending order based on AICc. The statistics provided in the Model List report consist of:
The likelihood function is a measure of how probable the observed data are, given the estimated model parameters. In a general sense, the higher the likelihood, the better the model fit. It follows that smaller values of -2Loglikelihood indicate better model fits.
AICc = -2loglikelihood + 2k + 2k(k + 1)/(n - k -1),
where k is the number of parameters in the model and n is the sample size. Smaller values of AICc indicate better model fits. The AICc penalizes the number of parameters, but less strongly than does the BIC.
BIC = -2loglikelihood + k ln(n),
where k is the number of parameters in the model and n is the sample size. Smaller values of BIC indicate better model fits. The BIC penalizes the number of parameters more strongly than does the AICc.
This provides analysis and scripting options that are available to all platforms. See the Using JMP book for more information about the options.