[source] # Least Squares Least Squares (or *quadratic* loss) is a function that computes the average squared error (MSE) between the target output given by the labels and the actual output of the network. It produces a smooth bowl-shaped gradient that is highly-influenced by large errors. $$ Least Squares = \sum_{i=1}^{D}(y_i-\hat{y}_i)^2 $$ ## Parameters This cost function does not have any parameters. ## Example ```php use Rubix\ML\NeuralNet\CostFunctions\LeastSquares; $costFunction = new LeastSquares(); ```