Wolfram Mathematica

« first day (1629 days earlier) ← previous day next day → last day (2852 days later) »

12:16 AM

Is there a way to fit this kind of data:

{{0, {1., 1., 1.}}, {1, {0., 0., 0.717647}}, {2, {0.00392157, 0., 0.862745}}, {3, {0., 0.00392157, 0.984314}}, {4, {0., 0.141176,  1.}}, {5, {0., 0.286275, 0.988235}}, {6, {0., 0.419608,  1.}}, {7, {0., 0.568627, 0.996078}}, {8, {0.00392157, 0.717647, 1.}}, {9,{0.00784314, 0.854902, 1.}}, {10, {0.00392157, 1., 1.}}, {11, {0.141176, 1., 0.866667}}, {12, {0.278431, 1., 0.709804}}, {13, {0.419608, 1., 0.576471}}}

I tried to do `Interpolation but it didn't work as expected

halirutan

@psimeson You realize that fitting a model and interpolating data are different things?

psimeson

I am trying to map the legend to the data mathematica.stackexchange.com/questions/119666/…

@halirutan I do understand but I am just trying every tools to understand that nonlinear scale

@halirutan I think mathematica can't do NonLinearFit do the data mentioned above, right? {1,1,1} ...are RGB

halirutan

@psimeson What about:

data = {{0, {1., 1., 1.}}, {1, {0., 0., 0.717647}}, {2, {0.00392157,
     0., 0.862745}}, {3, {0., 0.00392157, 0.984314}}, {4, {0.,
     0.141176, 1.}}, {5, {0., 0.286275, 0.988235}}, {6, {0., 0.419608,
      1.}}, {7, {0., 0.568627, 0.996078}}, {8, {0.00392157, 0.717647,
     1.}}, {9, {0.00784314, 0.854902, 1.}}, {10, {0.00392157, 1.,
     1.}}, {11, {0.141176, 1., 0.866667}}, {12, {0.278431, 1.,
     0.709804}}};
ip = Interpolation[data];

Then you can call values in the middle like ip[3.4]

psimeson

@halirutan btw the data shown there is no the full data... as can be noticed from the plot I can get more data to interpolation... I did that but the problem is that I have to get RGB value for each time.. but I do not get that when try interpolation

{{0, {1., 1., 1.}}, {1, {0., 0., 0.717647}}, {2, {0.00392157, 0.,
0.862745}}, {3, {0., 0.00392157, 0.984314}}, {4, {0., 0.141176,
1.}}, {5, {0., 0.286275, 0.988235}}, {6, {0., 0.419608,
1.}}, {7, {0., 0.568627, 0.996078}}, {8, {0.00392157, 0.717647,
1.}}, {9, {0.00784314, 0.854902, 1.}}, {10, {0.00392157, 1.,
1.}}, {11, {0.141176, 1., 0.866667}}, {12, {0.278431, 1.,
0.709804}}, {13, {0.419608, 1., 0.576471}}, {14, {0.568627,
0.996078, 0.427451}}, {35, {0.992157, 0.85098,
0.00392157}}, {96, {1., 0.137255, 0.00392157}}}

that's the full data

if = Interpolation[data, Method -> "Spline"]

halirutan

@psimeson Use your full data with Interpolation and try:

Plot[x, {x, 0, 96}, ColorFunction -> (RGBColor @@ ip[#] &),
 ColorFunctionScaling -> False, Filling -> Axis]

you get

psimeson

12:26 AM

RGBColor[if[34]]

halirutan

psimeson

that works but I think it's not representing the legend .. moreover when I try RGBColor[ip[34]] I do not get a color

halirutan

@psimeson Remember that ip[34] gives you a list of numbers, but RGBColor does not take a list as argument. It just needs a sequence of 3 numbers.

@psimeson There are two problems here. The first one is:

psimeson

This gives "RGBColor[ip[0]]" a color

halirutan

37 secs ago, by halirutan

@psimeson Remember that ip[34] gives you a list of numbers, but RGBColor does not take a list as argument. It just needs a sequence of 3 numbers.

@psimeson The second problem is that the numbers of ip[34] are not in the range [0,1].

That seems to be the reason that the RGBColor is not displayed as color-display

psimeson

12:31 AM

so, I think I have to do something else to get the true data out

halirutan

@psimeson See this:

psimeson

that seems to work.. I have never used clip

halirutan

@psimeson How about using InterpolationOrder->1?

This will prevent that the interpolation function is swinging into negative values or values greater 1.

psimeson

that works

halirutan

@psimeson Btw, it is completely not clear whether you really want a linear interpolation in RGB space. That is something you have to decide.

psimeson

12:34 AM

it seems like InterpolationOrder->1 is the way to go

I do not know how it legend scaled ... whatever gives the sensible result.. I have to go that way

@halirutan this seems to be the better representation

psimeson

12:51 AM

I think I am doing this in a reverse way. I need number out of colors rather than colors out of numbers.

3 hours later…

kirma

3:55 AM

Stanford Seminar - Stephen Wolfram of Wolfram Research

►

I suppose v11 was targeted for release one quarter after 10.4, but didn't quite make it...

2 hours later…

kirma

6:17 AM

Eh. Desktop Mathematica home edition has an annual payment license option, but it doesn't include upgrades. I was almost about to consider moving to annual licensing, but there's really no point in that case...

Michael Hale

7:11 AM

@kirma Oh? I thought it did. That's unfortunate.

kirma

Counterintuitively I started considering purchase of a pro license, though. On long run, it would be only roughly twice as expensive as upgrading home edition constantly! I'm not entirely certain if it makes sense considering that "long run" is easily a decade.

@MichaelHale At least that's the idea I get by carefully reading the alternatives. Cloud options surely auto-upgrade, but they're hardly enough for even hobbyist tasks... it's not uncommon to end up with computations taking hours and tens of gigabytes of RAM.

Basically annual pricing seems to be cost of having the cloud option, or an "I'm not sure if I want to commit to this product" pricing model just beyond free trial.

Apparently "starter" licenses that had lower price for the first year of standard (or pro, in contrast to home edition) licenses have disappeared from the list of offerings, although some popups still mention them.

1 hour later…

Szabolcs

8:21 AM

Does anyone understand why MaxExtraConditions behaves the way it does? mathematica.stackexchange.com/a/119951/12 (@kirma maybe you're interested)

@kirma That's okay. What's important is that whatever they release should be high quality and truly ready for release. I'd much rather see a polished release late than a beta-quality one early.

Szabolcs

8:41 AM

@J.M. I'm just giving you a warning that the golden ratio guy has been difficult in the past, on other sites ... meta.stackexchange.com/posts/280612/revisions maybe it's a good idea to keep the comments in check to avoid escalating the problem.