This example illustrates a log-linear model of independence, using data that contain structural zero frequencies as well as sampling (random) zero frequencies. In a population of six squirrel monkeys, ...

Model the data as a log-linear model with (the Poisson variance function) and where Yij= number of epileptic seizures in interval j tij= length of interval j The correlations between the counts are ...

Why yes. Suppose I take the same data from the pylab example and I imagine trying to add a linear function to represent that data. Here are two choices. Which is better? The red line or the blue one?