Thank you for the fine solutions. For my purpses the curve2 is the
best: It is a smooth curve and resutls close to the original monthly
temperature averages:
curve2: _7.6 _6.0 _1.7 4.3 11.4 17.2 19.6 17.3 11.7 5.1 _0.8 _5.1
original: _7.3 _6.6 _1.8 3.6 11.1 17.1 19.8 17.1 11.3
4.7 _1.3 _5.0
I think I can live with this difference. Thank you for the help and
teching fine J programming.
Anssi
At 23:22 8.6.2007, you wrote:
On 6/8/07, Brian Schott <[EMAIL PROTECTED]> wrote:
Neither am I an expert. Your plot solution looks
suspicious to me because the endpoints of each graph
"diverge" in the sense that for example the blue line is
arcing upward on both ends.
I took two different approaches for the end points.
In 'fit', I extended the data linearly, at the end points, then
discarded the extra data it produced. In 'curve', I did not
do so.
If you prefer, you might use
curve2=: 30 }. _30 }. [EMAIL PROTECTED]
in place of curve
Also, it's interesting to set arbitraryConstant to something
around 45, and compare the results from fit and curve.
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
