Non-linear regression

I have come across the documentations for regression but am a little confused. I wanted to obtain a non-linear regression model using the following: 'sea', 'sst' and 'at'. (See files attached). I wanted to check how sst and at affect sea levels. (something similar like: 'sea' = 'sst' + 'at')

I came across this sample code:

load carbig
tbl = table(Horsepower,Weight,MPG);
modelfun = @(b,x)b(1) + b(2)*x(:,1).^b(3) + ...
    b(4)*x(:,2).^b(5);
beta0 = [-50 500 -1 500 -1];
mdl = fitnlm(tbl,modelfun,beta0)

However, when I tried it on my data, I got a bad result. I would be grateful if someone could help me obtain the non-lin reg model for my data. If possible, could you send me the code too.

Thanks!

[Kevin Chng]

Hi, do you have the model?

modelfun = @(b,x)b(1) + b(2)*x(:,1).^b(3) + ...
           b(4)*x(:,2).^b(5);

or you need MATLAB to find the model for you?

[Keegan Carvalho]

@Kevin Chng I prefer if MATLAB finds the model. Since I do not know the type of model and equation for my analyses.

[Kevin Chng]

Take a look at this video https://www.mathworks.com/videos/curve-fitting-toolbox-overview-61198.html Maximun variable is 3.

If it does not suit you, you have to consider machine learning Link for reference

[Keegan Carvalho]

The link is not available. The number of variables is 3 so I used the code:

beta0 = [100 100 100];

But i still get an error:

Error using internal.stats.parseArgs (line 42) Wrong number of arguments.

Error in NonLinearModel.fit (line 1385) internal.stats.parseArgs(paramNames, paramDflts, otherArgs{:});

Error in fitnlm (line 99) model = NonLinearModel.fit(X,varargin{:});

[Kevin Chng]

Edited my comment. Sorry for the missing.

Answer 1

[Image Analyst]

I think you're being overly optimistic if you think that data can be modeled -- predicted by some analytical function.

data = readtable('ban1.csv')
years = data.year;
sea = data.sea;
sst = data.sst;
at = data.at;
subplot(2, 2, 1);
plot(years, sea, 'b-', 'LineWidth', 2);
grid on;
title('sea vs. year', 'FontSize', 20);
xlabel('Year', 'FontSize', 20);
ylabel('sea', 'FontSize', 20);

subplot(2, 2, 2);
plot(years, sst, 'b-', 'LineWidth', 2);
xlabel('Year', 'FontSize', 20);
grid on;
hold on;
plot(years, at, 'r-', 'LineWidth', 2);
grid on;
title('sst and at vs. year', 'FontSize', 20);
legend('sst', 'at', 'location', 'west');

subplot(2, 2, 3)
scatter(sst, sea, 'filled');
grid on;
title('sea vs. sst', 'FontSize', 20);
xlabel('sea', 'FontSize', 20);
ylabel('sst', 'FontSize', 20);

subplot(2, 2, 4)
scatter(sst, at, 'filled');
grid on;
title('at vs. sst', 'FontSize', 20);
xlabel('sea', 'FontSize', 20);
ylabel('at', 'FontSize', 20);