Hi, I have data of the form:
t x y trip
t1 x1+e y1+e A
t2 x2+e y2+e A
t3 x3+e y3+e B
t4 x4+e y4+e B
t5 x5+e y5+e B
... ... ... ...
where t is time and x/y are positions in space and e is a random error
term. Trips A/B/C/etc are entirely independent from each other. I'd like to
estimate the total distance traveled for each series (trip) of points
(there are actually a few thousand points in each trip, and thousands of
trips). A loess curve fits the data nicely, essentially eliminating the
error term (I realize not exactly but it is close enough for my purposes).
However I don't see an easy way to find arc length of a loess fit (without
using an inefficient loop of distance between t1,t2 + distance t2,t3, etc.)
Thanks for any advice.
[[alternative HTML version deleted]]
