SEMdeep (version 0.1.0)

getGradientWeight: Gradient Weight Approach for neural network variable importance

Description

The function computes the gradient matrix, i.e., the average conditional effects of the input variables w.r.t the neural network model, as discussed by Amesöder et al (2024).

Usage

getGradientWeight(object, thr = NULL, verbose = FALSE, ...)

Value

A list od two object: (i) a data.frame including the connections together with their weights, and (ii) the DAG with colored edges. If abs(W) > thr and W < 0, the edge is inhibited and it is highlighted in blue; otherwise, if abs(W) > thr and W > 0, the edge is activated and it is highlighted in red.

Arguments

object

A neural network object from SEMdnn() function.

thr

A threshold value to apply to gradient weights of input nodes (variables). If NULL (default), the threshold is set to thr=mean(abs(gradient weights)).

verbose

A logical value. If FALSE (default), the processed graph will not be plotted to screen.

...

Currently ignored.

Author

Mario Grassi

Details

The partial derivatives method calculates the derivative (the gradient) of each output variable (y) with respect to each input variable (x) evaluated at each observation (i=1,...,n) of the training data. The contribution of each input is evaluated in terms of both magnitude taking into account not only the connection weights and activation functions, but also the values of each observation of the input variables. Once the gradients for each variable and observation, a summary gradient is calculated by averaging over the observation units. Finally, the average weights are entered into a matrix, W(pxp) and the element-wise product with the binary (1,0) adjacency matrix, A(pxp) of the input DAG, W*A maps the weights on the DAG edges. Note that the operations required to compute partial derivatives are time consuming compared to other methods such as Olden's (connection weight). The computational time increases with the size of the neural network or the size of the data. Therefore, the function uses a progress bar to check the progress of the gradient evaluation per observation.

References

Amesöder, C., Hartig, F. and Pichler, M. (2024), ‘cito': an R package for training neural networks using ‘torch'. Ecography, 2024: e07143. https://doi.org/10.1111/ecog.07143

Examples

# \donttest{
if (torch::torch_is_installed()){

# load ALS data
ig<- alsData$graph
data<- alsData$exprs
data<- transformData(data)$data

dnn0 <- SEMdnn(ig, data, train=1:nrow(data), cowt = FALSE,
			#loss = "mse", hidden = 5*K, link = "selu",
			loss = "mse", hidden = c(10, 10, 10), link = "selu",
			validation = 0, bias = TRUE, lr = 0.01,
			epochs = 32, device = "cpu", verbose = TRUE)

res<- getGradientWeight(dnn0, thr=NULL, verbose=TRUE)
table(E(res$dag)$color)
}
# }