fitlm and exponential model help

Hi

I have a set of data, which is voltage from a sensor aganist temperture.

As temperture increase the voltage output from the sensor will drop, even though the sensing enviroment in the same.

Therefore i need to be able to calulate the correction factor so the output is very close to when sensor is expose to 20 degrees.

I am using fitlm, but when i use exponential i am getting errors, which is not then when i use Quadratic

This is what i am using

no2Model = fitlm(iot_temperature, voltage_differnce_no2, "exponential");

Is this the correct command to use?

[Mathieu NOE]

hello

can you share the data + code ?

Answer 1

[Star Strider]

The "exponential" option for modelspec does not appear to exist.

The fitnlm function is likely more appropriate. That is more likely to produce the result you want.

Example —

t = linspace(0, 10, 25);
y = 3*exp(-0.25*t) + 0.5*randn(size(t)) + 1;

fcn = @(b,x) b(1).*exp(b(2).*x) + b(3);

B0 = randn(3,1);
mdl = fitnlm(t,y,fcn,B0)

mdl = 
Nonlinear regression model:
    y ~ b1*exp(b2*x) + b3

Estimated Coefficients:
          Estimate      SE        tStat       pValue  
          ________    _______    _______    __________

    b1     2.7569     0.44251     6.2301    2.8587e-06
    b2    -0.2526     0.12197    -2.0711      0.050281
    b3     1.0658     0.48102     2.2157      0.037369


Number of observations: 25, Error degrees of freedom: 22
Root Mean Squared Error: 0.519
R-Squared: 0.696,  Adjusted R-Squared 0.669
F-statistic vs. constant model: 25.2, p-value = 2.01e-06

[yfit,yci] = predict(mdl,t(:));


figure
hp{1} = plot(t, y, '.', 'DisplayName','Data');
hold on
hp{2} = plot(t, yfit, '-r', 'DisplayName','Model Fit');
hp{3} = plot(t, yci, ':r', 'DisplayName','±95% CI');
hold off
grid
xlabel('t')
ylabel('y')
legend([hp{1},hp{2},hp{3}(1)], 'Location','best')

.