How to build a polynomial regression model with multiple independent variables using Python?

Question

Python

How to build a polynomial regression model of the second degree with several independent variables (from 2x to 4x) in such a way that you can see the equation itself and the data to assess the adequacy of the model? The data array includes 50 dependent variables and 4 independent variables for each of them.

Answer 1

There are many options.
sklearn.preprocessing.PolynomialFeatures is the time.
You can manually convert a multivariate polynomial model to a multivariate linear one, and then run
sklearn.LinearRegression()
You can use the more general variant
scipy.optimize.curve_fit() , and since regression is essentially a minimization problem, any function can be fitted here. Moreover, scipy.optimize has many different numerical optimization methods that can be used to obtain a regression model, including.
In TensorFlow, you can use the
GradientDescentOptimizer
. But I recommend that you first understand the task and terminology. Your phrase
"50 dependent variables" - at first plunged me into a stupor.