Quantum KRR over GAT node embeddings (Pilar 6 x Pilar 10)

Thin composition: take the GAT embedding matrix, PCA-reduce to n_pcs, feed to quantum_krr_fit(). Returns a combined fit object that predict() unwraps by projecting new embeddings through the stored rotation.

Usage

qf_krr_on_gat_embeddings(gnn_fit, y, n_pcs = 6L, reps = 2L, lambda = 0.5)

Arguments

gnn_fit

An edaphos_gnn_gat fit.

y

Numeric response (one per training node).

n_pcs

Integer; number of PCs (= qubits). Default 6L.

reps

ZZFeatureMap repetitions. Default 2L.

lambda

Ridge regulariser. Default 0.5.

Value

An edaphos_qf_krr_gat fit.