Sum-of-squared-error loss for continuous outcomes

Formula: L(w) = 0.5 * sum_i sum_j (yhat_ij - y_ij)^2

Usage

Ssello(Y, Yhat)

Arguments

Y

Observed values, one row per record; a flat vector is read as one column.

Yhat

Predicted values, same shape as the observed matrix.

Value

List with loss, mean_loss, n, L.

References

Montesinos Lopez, Montesinos Lopez and Crossa (2022), Multivariate Statistical Machine Learning Methods for Genomic Prediction, Springer, doi:10.1007/978-3-030-89010-0. Chapter 10, Sect. 10.7, pp. 400-403. Read from the chapter PDF, not recalled. The book notes that dividing by two is for convenience in the backpropagation gradient, and that it is also common to report the SSE divided by n times L; mean_loss is that per-cell value.

Examples

V <- c(1, 2, 3, 4, 5, 6, 7, 8)
Ssello(V, V)
#> $loss
#> [1] 0
#> 
#> $mean_loss
#> [1] 0
#> 
#> $n
#> [1] 8
#> 
#> $L
#> [1] 1
#> 
#> $method
#> [1] "SSE loss, MVSML Sect. 10.7.1"
#> 
#> attr(,"class")
#> [1] "morie_rich_result" "list"