Model fitting for classification (2): probabilistic version of problem statement

Antonio Sala, UPV

Difficulty: ** ,       Relevance: ,      Duration: 15:41

*Enlace a Spanish version

Summary:

This video presents the “probabilistic” version of the problem of fitting $𝜃$ in a model $f(x,𝜃)$ to fit pairs of training data $(x_i,y_i)$ with $y_i\in \{0,1\}$, that is, binary classification. The “deterministic” version is discussed in the video [clasifintr1EN], previous to this one, which you are advised to watch.

The objective here is for the model to estimate the probability of having a certain output at 1 given $x$... the simplest version is to maximize the probability of the samples (maximum likelihood estimation) by adjusting $𝜃$. There may be more formal Bayesian approaches, out of the scope of this very short introduction to the topic.

*Link to my [ whole collection] of videos in English. Link to larger [ Colección completa] in Spanish.