nn_option_pricing.noisy_svr_experiment

Support Vector Regression benchmark with noisy Black-Scholes targets.

Functions

run_noisy_svr_benchmark(config[, output_dir])	Run the noisy-target SVR benchmark over all levels and seeds.
run_single_noisy_svr_benchmark(seed, ...)	Train SVR on one noisy target dataset and evaluate clean/noisy metrics.
summarize_noisy_svr_runs(runs)	Aggregate noisy SVR metrics by noise level.

Classes

NoisySVRBenchmarkConfig([n_samples, seeds, ...])

Configuration for reduced-scale noisy-target SVR experiments.