enh: For the inspection of linear models, add statistical tests to the coefficients

[sylvaincom]

Is your feature request related to a problem? Please describe.

For feature importance on linear models, we display the coefficients (see #1339 and the example on the doc). This might not be enough.

Describe the solution you'd like

Add statistical significance

as in R

http://r-statistics.co/Linear-Regression.html

summary(linearMod)  # model summary
#> Call:
#> lm(formula = dist ~ speed, data = cars)
#> 
#> Residuals:
#>     Min      1Q  Median      3Q     Max 
#> -29.069  -9.525  -2.272   9.215  43.201 
#> 
#> Coefficients:
#>             Estimate Std. Error t value Pr(>|t|)    
#> (Intercept) -17.5791     6.7584  -2.601   0.0123 *  
#> speed         3.9324     0.4155   9.464 1.49e-12 ***
#> ---
#> Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#> 
#> Residual standard error: 15.38 on 48 degrees of freedom
#> Multiple R-squared:  0.6511, Adjusted R-squared:  0.6438 
#> F-statistic: 89.57 on 1 and 48 DF,  p-value: 1.49e-12

as in Python, statsmodels

I do not believe scikit-learn does it.
In the the statsmodels package, there are statistical tests associated with each coefficient:

https://www.statsmodels.org/stable/regression.html

Describe alternatives you've considered, if relevant

This is P2. If we show so many statistical tests, we would need to explain our users what each test does, which might not be desirable. Maybe just show the stars associated to the significance of a feature?

Additional context

No response