A classmate and I are doing a projet on risk management. To cut a long story short, we need check whether our insurance data ( US$ losses caused by natural catastrophes) follow a poisson process. We used the chi2gof test as follows:
where 0.2596 is the lambda we estimated by maximum likelihood. However, this test has a default value of significance of 5% and apparently one of the catastrophes doesn't follow a poisson process at such a level. Therefore, we would like to know whether there exists another test for a poisson process that allows us to choose the significance level.
Someone suggested using the dgt test for a poisson process but I can't find evidence of it. Does anyone know if it really exists (maybe under another name)?