MATLAB: Interpolation, sigmoid curve

interpolationMATLABproblem

Hi, I'm not very good on matlab, som maybe this is a kind of basic question.

I have 140 values from 140 years of coal mining. There is a maximum limit of how much coal that can be extracted from the mine. My task is to make a prognosis for the next 60 years. It's suppose to look lika a sigmoind curve (an S). I have been able to make a sigmoid curve based on the different values from the years i already have. But my problem is how i'm soppose to interpolate these values into the future?

I have tried to use interp1, but i my result is not that good, it's ether going straight to the roof or below 0. So I would be very grateful for any help that I can get!

Best Answer

you mean extrapolate into the future, not interpolate ... If you have Optimization Toolbox, you can try fitting the generalized logistic equation to your data (<http://en.wikipedia.org/wiki/Generalised_logistic_curve>) using 'lsqnonlin'

Related Solutions

MATLAB: Lagrange interpolation, .m

pointx and pointy are two vectors of data values, x is a vector of points where you want to interpolate. For example:

x = 0:10;
y = x.^2;
xx = linspace(0,10);
yy = lagrange(xx,x,y);
plot(x,y,'o',xx,yy,'.')

As an aside, with no offense intended to Calzino, there are other options available for interpolation. Firstly, of course, interp1 is a standard MATLAB function, with options for linear, cubic spline, and PCHIP interpolation. Cleve Moler (aka The Guy Who Wrote MATLAB) also has a Lagrange interpolation function available for download.

MATLAB: Suitable Regression or Curve fitting for vertical scattered Data

I guess you can try regression using artificial neural network (ANN):

http://www.mathworks.com/help/nnet/function-approximation-and-nonlinear-regression.html

Or you can go for some techniques in statistics and machine learning toolbox...

http://www.mathworks.com/help/stats/functionlist.html#btq_uq5

Best Answer

Related Solutions

MATLAB: Lagrange interpolation, .m

MATLAB: Suitable Regression or Curve fitting for vertical scattered Data

Related Question