Weights and intercept values

[eddyhub]

Hello , i have three questions :

• how do you decide the weights associated with the feature values ? it could be any number
• i didn’t understand what intercept mean and also as above how you decide intercepts value
• how the alpha value in the scatterplot works? it is explained as proportional opacity of the points. but what is the meaning of having data points with different opacity ?

[ArturoAmorQ]

Hello @eddyhub

@eddyhub

how do you decide the weights associated with the feature values ? it could be any number

This notebook is only to build up the intuitions of a good or not-so-good fit. The weights and intercepts are arbitrary and you can play with those parameters to try to make a better fit by hand.

@eddyhub

i didn’t understand what intercept mean

It is the point where the fit intercepts the “Y” axis. Provided a straight line with equation y = mx + b, the intercept is b.

@eddyhub

what is the meaning of having data points with different opacity ?

It’s only purpose is for easing visualization. It follows the notion of darker equals more overlapping points.