Re: [R] nls problem: singular gradient
On 07/12/2012 01:39 AM, Duncan Murdoch wrote: On 12-07-11 2:34 PM, Jonas Stein wrote: Take a look at the predicted values at your starting fit: there's a discontinuity at 0.4, which sure makes it look as though overflow is occurring. I'd recommend expanding tanh() in terms of exponentials and rewrite the prediction in a way that won't overflow. Duncan Murdoch Hi Duncan, Thank you for your suggestion. I wrote a function mytanh and nls terminates a bit later with another error message: Error in nls(data = dd, y ~ 1/2 * (1 - mytanh((x - ttt)/1e-04) * exp(-x/tau2)), : number of iterations exceeded maximum of 50 How can i fix that? Kind regards, Jonas R CODE STARTS HERE === mytanh- function(x){ return(x - x^3/3 + 2*x^5 /15 - 17 * x^7/315) } That looks like it would overflow as soon as abs(x-ttt) got large, just like the original. You might be able to fix it by following the advice I gave last time, or maybe you need to rescale the parameters. In most cases optimizers work best when the uncertainty in the parameters is all on the same scale, typically around 1. I am not shure what you mean with rescale paramaeters, but i changed ttt and tau2 to 1 but nls still fails. Do you mean i can only use functions with tau2 and ttt close to 1? Is there a better fit function then nls for R? Even origin can find the parameters without any problems. nlsfit - nls(data=dd, y ~ 1/2 * ( 1- mytanh((x - ttt)/0.0001) * exp(-x / tau2) ), start=list(ttt=1, tau2=1) , trace=TRUE, control = list(maxiter = 100)) -- Jonas Stein n...@jonasstein.de __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] nls problem: singular gradient
On Jul 11, 2012, at 20:34 , Jonas Stein wrote: Take a look at the predicted values at your starting fit: there's a discontinuity at 0.4, which sure makes it look as though overflow is occurring. I'd recommend expanding tanh() in terms of exponentials and rewrite the prediction in a way that won't overflow. Duncan Murdoch Hi Duncan, Thank you for your suggestion. I wrote a function mytanh and nls terminates a bit later with another error message: Error in nls(data = dd, y ~ 1/2 * (1 - mytanh((x - ttt)/1e-04) * exp(-x/tau2)), : number of iterations exceeded maximum of 50 How can i fix that? Kind regards, Jonas R CODE STARTS HERE === mytanh - function(x){ return(x - x^3/3 + 2*x^5 /15 - 17 * x^7/315) } Ouch! Your original had something that was very nearly a Heaviside function, i.e. a function that instantaneously switches from 1 to 0 at x=ttt. Replace that with a polynomial and I bet that your fitted curves has no resemblance to the original. Change-point estimation is notoriously tricky because the sum of squares changes discontinuously as points cross to the other side and there will be regions where the objective function is constant, because you can move the change point and still have the same sets of points on each side of. Your function tries to remedy this by using a soft threshold, but it is still quite an abrupt change: if you put ttt in the middle of an interval of length 0.001. Then abs((x-ttt)/0.001) will be at least 5 and tanh(+5) [1] 0.092 Furthermore, the derivative of tanh at that point is roughly (tanh(4.999) - tanh(5.001))/0.002 [1] -0.0001815834 I.e. moving the change point yields only a very small change in the fitted value at the neighboring x values and hardly a change at all at any other point. This is where your singular gradient comes from. Pragmatically, you could try a larger value than 0.0001, but I suspect it would be wise to supplement any gradient technique with a more direct search procedure. Overall, I'd say that you need a bit more Fingerspitzgefühl with this sort of optimization problem. Try this, for instance: x - seq(0,1,.001) y - (x 1/pi) + rnorm(x, sd=.01) plot(x,y) a - seq(0,1,0.01) S - sapply(a, function(a) sum((y - (xa))^2)) # SSD plot(a,S) # Looks easy enough? a - seq(.31,.32,0.0001) # Now zoom in S - sapply(a, function(a) sum((y - (xa))^2)) plot(a,S) # Oops, try soft threshold S - sapply(a, function(a) sum((y - 0.5*(tanh((x-a)/.0001)+1))^2)) plot(a,S) # Umm, maybe soften a bit more S - sapply(a, function(a) sum((y - 0.5*(tanh((x-a)/.001)+1))^2)) plot(a,S) -pd t - seq(0,1,0.001) t0 - 0.5 tau1 - 0.02 yy - 1/2 * ( 1- tanh((t - t0)/0.0001) * exp(-t / tau1) ) + rnorm(length(t))*0.001 plot(x=t, y=yy, pch=18) dd - data.frame(y=yy, x=t) nlsfit - nls(data=dd, y ~ 1/2 * ( 1- mytanh((x - ttt)/0.0001) * exp(-x / tau2) ), start=list(ttt=0.5, tau2=0.02) , trace=TRUE) R CODE ENDS HERE === -- Jonas Stein n...@jonasstein.de __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- Peter Dalgaard, Professor Center for Statistics, Copenhagen Business School Solbjerg Plads 3, 2000 Frederiksberg, Denmark Phone: (+45)38153501 Email: pd@cbs.dk Priv: pda...@gmail.com __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] nls problem: singular gradient
Duncan has given a indication of why nls() has troubles, and you have found a way to work around the problem partially. However, you may like to try nlmrt (from R-forge project R-forge.r-project.org/R/?group_id=395 It is intended to be very aggressive in finding a solution, and also to deal with small residual problems that are really not statistical in nature i.e., nonlinear least squares but not nonlinear regression. Note that there is a function wrapnls() if you want the nls() output structure which is very useful for modeling. nlmrt::nlxb is closer to an optimization method. I'd be interested to know if you find the solution found by nlmrt is useful in your context. -- require(nlmrt) nlsfit2- nlxb(data=dd, y ~ 1/2 * ( 1- tanh((x - ttt)/smallc) * exp(-x / tau2) ), start=list(ttt=0.4, tau2=0.1) , trace=TRUE) -- Best, JN On 07/12/2012 06:00 AM, r-help-requ...@r-project.org wrote: From: Jonas Stein n...@jonasstein.de To: r-h...@stat.math.ethz.ch Subject: [R] nls problem: singular gradient Message-ID: e8h0d9-ao4@news.jonasstein.de Content-Type: text/plain Why fails nls with singular gradient here? I post a minimal example on the bottom and would be very happy if someone could help me. Kind regards, ### # define some constants smallc - 0.0001 t - seq(0,1,0.001) t0 - 0.5 tau1 - 0.02 # generate yy(t) yy - 1/2 * ( 1- tanh((t - t0)/smallc) * exp(-t / tau1) ) + rnorm(length(t))*0.01 # show the curve plot(x=t, y=yy, pch=18) # prepare data dd - data.frame(y=yy, x=t) nlsfit - nls(data=dd, y ~ 1/2 * ( 1- tanh((x - ttt)/smallc) * exp(-x / tau2) ), start=list(ttt=0.4, tau2=0.1) , trace=TRUE) # get error: # Error in nls(data = dd, y ~ 1/2 * (1 - tanh((x - ttt)/smallc) * exp(-x/tau2)), : # singular gradient -- Jonas Stein n...@jonasstein.de __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] nls problem: singular gradient
Why fails nls with singular gradient here? I post a minimal example on the bottom and would be very happy if someone could help me. Kind regards, ### # define some constants smallc - 0.0001 t - seq(0,1,0.001) t0 - 0.5 tau1 - 0.02 # generate yy(t) yy - 1/2 * ( 1- tanh((t - t0)/smallc) * exp(-t / tau1) ) + rnorm(length(t))*0.01 # show the curve plot(x=t, y=yy, pch=18) # prepare data dd - data.frame(y=yy, x=t) nlsfit - nls(data=dd, y ~ 1/2 * ( 1- tanh((x - ttt)/smallc) * exp(-x / tau2) ), start=list(ttt=0.4, tau2=0.1) , trace=TRUE) # get error: # Error in nls(data = dd, y ~ 1/2 * (1 - tanh((x - ttt)/smallc) * exp(-x/tau2)), : # singular gradient -- Jonas Stein n...@jonasstein.de __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] nls problem: singular gradient
On 11/07/2012 11:04 AM, Jonas Stein wrote: Why fails nls with singular gradient here? I post a minimal example on the bottom and would be very happy if someone could help me. Take a look at the predicted values at your starting fit: there's a discontinuity at 0.4, which sure makes it look as though overflow is occurring. I'd recommend expanding tanh() in terms of exponentials and rewrite the prediction in a way that won't overflow. Duncan Murdoch Kind regards, ### # define some constants smallc - 0.0001 t - seq(0,1,0.001) t0 - 0.5 tau1 - 0.02 # generate yy(t) yy - 1/2 * ( 1- tanh((t - t0)/smallc) * exp(-t / tau1) ) + rnorm(length(t))*0.01 # show the curve plot(x=t, y=yy, pch=18) # prepare data dd - data.frame(y=yy, x=t) nlsfit - nls(data=dd, y ~ 1/2 * ( 1- tanh((x - ttt)/smallc) * exp(-x / tau2) ), start=list(ttt=0.4, tau2=0.1) , trace=TRUE) # get error: # Error in nls(data = dd, y ~ 1/2 * (1 - tanh((x - ttt)/smallc) * exp(-x/tau2)), : # singular gradient __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] nls problem: singular gradient
Take a look at the predicted values at your starting fit: there's a discontinuity at 0.4, which sure makes it look as though overflow is occurring. I'd recommend expanding tanh() in terms of exponentials and rewrite the prediction in a way that won't overflow. Duncan Murdoch Hi Duncan, Thank you for your suggestion. I wrote a function mytanh and nls terminates a bit later with another error message: Error in nls(data = dd, y ~ 1/2 * (1 - mytanh((x - ttt)/1e-04) * exp(-x/tau2)), : number of iterations exceeded maximum of 50 How can i fix that? Kind regards, Jonas R CODE STARTS HERE === mytanh - function(x){ return(x - x^3/3 + 2*x^5 /15 - 17 * x^7/315) } t - seq(0,1,0.001) t0 - 0.5 tau1 - 0.02 yy - 1/2 * ( 1- tanh((t - t0)/0.0001) * exp(-t / tau1) ) + rnorm(length(t))*0.001 plot(x=t, y=yy, pch=18) dd - data.frame(y=yy, x=t) nlsfit - nls(data=dd, y ~ 1/2 * ( 1- mytanh((x - ttt)/0.0001) * exp(-x / tau2) ), start=list(ttt=0.5, tau2=0.02) , trace=TRUE) R CODE ENDS HERE === -- Jonas Stein n...@jonasstein.de __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] nls problem: singular gradient
On 12-07-11 2:34 PM, Jonas Stein wrote: Take a look at the predicted values at your starting fit: there's a discontinuity at 0.4, which sure makes it look as though overflow is occurring. I'd recommend expanding tanh() in terms of exponentials and rewrite the prediction in a way that won't overflow. Duncan Murdoch Hi Duncan, Thank you for your suggestion. I wrote a function mytanh and nls terminates a bit later with another error message: Error in nls(data = dd, y ~ 1/2 * (1 - mytanh((x - ttt)/1e-04) * exp(-x/tau2)), : number of iterations exceeded maximum of 50 How can i fix that? Kind regards, Jonas R CODE STARTS HERE === mytanh- function(x){ return(x - x^3/3 + 2*x^5 /15 - 17 * x^7/315) } That looks like it would overflow as soon as abs(x-ttt) got large, just like the original. You might be able to fix it by following the advice I gave last time, or maybe you need to rescale the parameters. In most cases optimizers work best when the uncertainty in the parameters is all on the same scale, typically around 1. Duncan Murdoch t- seq(0,1,0.001) t0- 0.5 tau1- 0.02 yy- 1/2 * ( 1- tanh((t - t0)/0.0001) * exp(-t / tau1) ) + rnorm(length(t))*0.001 plot(x=t, y=yy, pch=18) dd- data.frame(y=yy, x=t) nlsfit- nls(data=dd, y ~ 1/2 * ( 1- mytanh((x - ttt)/0.0001) * exp(-x / tau2) ), start=list(ttt=0.5, tau2=0.02) , trace=TRUE) R CODE ENDS HERE === __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.