nn_option_pricing.svr

Support Vector Regression benchmark utilities.

The neural network is the main surrogate model in the project. This module provides a reduced-scale Support Vector Regression baseline, useful as a classical machine-learning comparison without changing the Black-Scholes function-approximation research question.

Functions

run_single_svr_benchmark(seed, config)	Train and evaluate one SVR benchmark run for a single random seed.
run_svr_benchmark(config[, output_dir])	Run the SVR benchmark over all configured seeds.
summarize_svr_runs(runs)	Aggregate per-seed SVR benchmark metrics into mean/std summaries.

Classes

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

Configuration for the reduced-scale SVR benchmark.