How does one get polynomial features like in python but for MatLab?

Is there code that gets the polynomial features of a vector like in:

http://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.PolynomialFeatures.html

e.g. for 1D 1,x,x^2 for degree 2. Ideally like the link I posted.

Ideally, is there no built in function in matlab that does this for me?

Answer 1

[Sanjana Ramakrishnan]

There is not a direct way of creating a polynomial matrix in MATLAB. Refer the below link for information on polynomial matrices:

https://en.wikipedia.org/wiki/Polynomial_matrix

However, you can create a vector of polynomial coefficients in MATLAB to represent a polynomial: https://www.mathworks.com/help/matlab/math/create-and-evaluate-polynomials.html

You can then append the vectors vertically to form a polynomial matrix

[Brando Miranda]

what about fitting polynomials?

Answer 2

[Steven Lord]

For one variable, I think you want the vander function.

For multiple variables, use the x2fx function from Statistics and Machine Learning Toolbox. If you want terms with a degree greater than 2 (like x1.^2.*x2, with a degree of 2+1 = 3) you'd have to build the model input rather than using one of the predefined options like 'quadratic'. I believe there are some discussions on Answers and some files on the File Exchange for generating all combinations of N numbers that add up to a particular value that you could use to help you build the model input in that case.