Data Science

How to derive class probability from e.g. a linear regression?

- points that are far away (above) from the discriminant function are more likely to belong to the upper class (e.g. prob: 0.97)
- whereas points that are far away (below) from the discriminant function are more likely to belong to the lower class (e.g. prob 0.12)
- points close to the discriminant function are difficult to predict (e.g. prob 0.51)

Diskussion