bug in kalman smoother when k=0?

[rBatt]

https://github.com/GLEON/LakeMetabolizer/blob/master/R/metab.kalman.R#L295

The commit (ryan's) that created the bug is here

Resulted from trying to handle k=0. That's a bug, right??

@lawinslow

@cbuelo You said you'd had an issue with smoothing the oxygen, right? It might be that the earlier versions I tested hadn't yet tried to handle k=0, I'd have to check.

Note that this only affects the smoother of DO, not the estimation of metabolism.

[cbuelo]

Yeah if I remember correctly the returned "smooth" DO time series had big, random-looking jumps in it.

[rBatt]

That's probably due to dividing by a small fraction in the smoother, but
I'm not sure.

These issues may be unrelated. However, I'd be interested to see if your
smoothing problem was solved once this bug fix is implemented.

On Mon, Apr 27, 2015 at 9:20 AM, Cal [email protected] wrote:

Yeah if I remember correctly the returned "smooth" DO time series had big,
random-looking jumps in it.

—
Reply to this email directly or view it on GitHub
#92 (comment)
.

[rBatt]

This is now related to #126