gmm accuracy plot using the r script

for 6–100

Those numbers are pretty bad, right?

y axis is accuracy so max is 1 I think

but I don’t know how much it should change with time. If there is no change with sweeps then the benefit of running more sweeps doesn’t show up

I’m guessing “plots.pdf” is 3–100?

For both of these plots, how many data-sets (i.e., observed 100 reals) and how many random trials per data-set (including random initialization of inferred classifications) did you average across?