Services & Resources / Wolfram Forums / MathGroup Archive
-----

MathGroup Archive 2011

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: weibull plot on weibull scaled paper

  • To: mathgroup at smc.vnet.net
  • Subject: [mg116586] Re: weibull plot on weibull scaled paper
  • From: DrMajorBob <btreat1 at austin.rr.com>
  • Date: Mon, 21 Feb 2011 04:19:36 -0500 (EST)

ProbabilityScalePlot often throws a FindRoot::jsing error with random  
data, and I wonder why.

I'm also trying to see the relationship between these two graphs:

dist = WeibullDistribution[7, 200];
Quiet@ProbabilityScalePlot[data, "Weibull"]
QuantilePlot[data, dist]

Bobby

On Sun, 20 Feb 2011 04:25:48 -0600, Bob Hanlon <hanlonr at cox.net> wrote:

>
> Just label the y-axis with values of the CDF that correspond to the  -  
> Log[1 - CDF[dist, x]] values
>
> dist = WeibullDistribution[7, 200];
>
> LogLogPlot[-Log[1 - CDF[dist, x]], {x, 70, 300},
>  Frame -> True, Axes -> False,
>  FrameTicks -> {Automatic, {-Log[1 - #],
>       100 #} & /@
>     ({1, 5, 10, 25, 50, 75, 90, 99}/100)},
>  PlotRange -> (-Log[1 - #] & /@ {0.005, 0.999})]
>
>
> Bob Hanlon
>
> ---- Fred Schenkelberg <fms at garlic.com> wrote:
>
> =============
> Hi Bob,
>
> Thanks for the response.
>
> The loglog plot is similar to the classic weibull scales plotted when  
> using ProbabilityScalePlot, yet isn't the same. For the CDF the vertical  
> scale should go from near zero to near 100% (or 1 by some authors)
>
> When using just,
> ProbabilityScalePlot[data, "Weibull"]
>
> from your example - the scales loo right.
>
> And, I'm pretty sure that I can update the output to include labels,  
> classic grid lines, etc.
>
> The problem is that I've not been able to find a way to plot the  
> 'straight line' representing the known distribution on the weibull  
> probability paper scales, without having to use RandomVariate and fit a  
> line.  If that is the only way to get to the probability scaled paper,  
> then, is it possible to plot only the fitted line or reference and not  
> the data points?
>
>
> In my case, I have the weibull parameters, and not the data.
>
> It's one of those problems that should be easy....
>
> any ideas?
>
> cheers,
>
> Fred
>
>
> On Feb 19, 2011, at 5:45 AM, Bob Hanlon wrote:
>
>
> The distribution is just a straight line
>
> dist = WeibullDistribution[7, 200];
>
> llp = LogLogPlot[-Log[1 - CDF[dist, x]], {x, 70, 300},
>  Frame -> True, Axes -> False]
>
> For comparison
>
> data = RandomVariate[dist, 300];
>
> Show[llp, ProbabilityScalePlot[data, "Weibull"]]
>
>
> Bob Hanlon
>
> ---- Frederick <fms at garlic.com> wrote:
>
> =============
> Hi,  Is it possible to plot
>
> Plot[Evaluate[CDF[WeibullDistribution[7, 200], x]], {x, 0, 300}]
>
> using the classic weibull scaled paper, as available in
> ProbabilityScalePlot?
>
> and, without having to plot and/or show data dots....
>
> Looking for just a smooth plot of the known distribution on the
> classic weibull scaled graph
>
> thanks,
>
> Fred
>
>


-- 
DrMajorBob at yahoo.com


  • Prev by Date: Vector Runge-Kutta ODE solver with compilation?
  • Next by Date: Re: Color grid with x and y args to visualize
  • Previous by thread: Re: weibull plot on weibull scaled paper
  • Next by thread: Re: weibull plot on weibull scaled paper