Table 1.

The components of the Laplace approximation to the marginal likelihood and the corresponding approximate posterior model probabilities under two priors for the number of classes: a discrete uniform prior and a truncated Poisson prior

	Number of classes
	1	2	3	4
Number of parameters	9	19	28	37
log likelihood	–3571.829	–3501.31	–3489.768	–3489.751
$(1∕2)\log \mid \widehat{\Sigma }\mid$	–24.51	–44.07	–55.84	–78.22
d/2 log (2π)	8.27	17.46	25.73	34.00
P(m|y), uniform prior	0	0	1	0
P(m|y), truncated Poisson prior	0	0	1	0