prior define upper and lower on y

[olesendan]

Hi.

I'm trying to learn, how to use your package.

I'm messing around with priors, and would like to set bounds on y. Is this possible?
I think i can set cp_i = "T(MINX, MAXX)". Is it possible to do something like cp_i = T(y >= 1, y<=2)

Greetings from denmark
Dan Olesen

[lindeloev]

Hi Dan,

Sorry for missing your issue! Priors only apply to inferred parameters. If y is observed, it has no prior. cp_[some-number-here] is the parameter representing the (inferred) change point location. Say you wanted to put a bounded normal prior on the first change point:

mcp(segments, data, prior = list(cp_1 = "dnorm(10, 3) T(MINX, 20)"))

If you really want "bounds" on y, maybe that's a preprocessing step where you filter away "bad" values, like y = y[y > 0 & y < 10]. Maybe I'm answering the wrong question. If so, could you be a bit more specific?