Notation

Terminology:

Features: Independent variables used for predictive modeling
Outcome: Dependent variable
Feature matrix: Matrix of independent variables with samples along columns and features along rows
Example: One sample (column) from feature matrix

Notation:

$X$: Feature matrix containing several samples and one or more independent variables
$y$: Vector of outcomes - can sometimes be a one-hot matrix
$X^{(i)}$: $i^{th}$ sample/example/column in matrix X; matrix X includes a column with value 1 throughout to accommodate the intercept term in weights unless specified otherwise
$y^{(i)}$: $i^{th}$ value in vector y
$W$: Weight matrix of true model (unknown); includes the intercept unless specified otherwise
$\hat{W}$: Estimated weight matrix of the model
$N_{train}$: Number of samples in training set
var: Variance/covariance operator on a series or matrix
sd: Standard deviation operator on a series

Abbreviations: