Formal definition of Machine Learning given Classification as Example

169 Views Asked by At

I have tried to write some easy but formal definition of a machine learning model using a classification task as example and I need some review.

The goal of machine learning is:

$arg min_{\Theta}(c(\textbf y,\hat{\textbf y})|o)$

$c(.) \in C$ = Cost Function of a set of Cost Functions$

$f(.) \in F$ = Function (Machine Learning Model) of a set of Functions$

$o(.) \in O$ = Optimizer of a set of Optimizers$

$\hat{y} = f(\textbf x,\Theta) =$ Predicted Value

$y =$ True Value

$\textbf x \in \mathbb{R}^{d}$ = d-dimensional Datapoint$

$\Theta \in \mathbb{R}^{m}$ = m-dimensional Parameter Vector$

Here, given the dataset $\mathbf X$ for a chosen Classifier $f(.) \in F$ the Optimizer $o(.) \in O$ is used to find the Parameter Vector $\Theta \in \mathbb{R}^{m}$ for the $f(.) \in F$ such that the Cost function $c(.) \in C$ is minimized.

I am happy for any comments!