Building machine learning models

One of the most simple machine learning models we can construct to make a forecast of future behaviors is linear regression, which reduces the residual sum of squares between the targets observed in the dataset and the targets anticipated by the linear approximation, fitting a linear model using coefficients.

This is simply ordinary least squares or non-negative least squares wrapped in a predictor object from the implementation perspective.

We can implement this really simply by using the LinearRegression class in Sklearn:

from sklearn.linear_model import LinearRegression
from sklearn.datasets import load_diabetes
data_reg = load_diabetes()
x,y = data_reg['data'],data_reg['target']
reg = LinearRegression().fit(x, y)
reg.score(x, y)

Figure 2.24: Model regression score

The preceding code will fit a linear regression model to our data and print the score of our data.

We can also print the coefficients...