The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

CRAFT

CRAFT provides Conditional Regime Analog Forecasting with Trajectories tools for multivariate time series. It builds lag and lead trajectory embeddings, detects SVD-based changepoint regimes, estimates transition probabilities, samples plausible future-trajectory analogues, and fits smooth empirical forecast distributions.

Installation

After CRAN release:

install.packages("CRAFT")

For local development from this folder:

install.packages(".", repos = NULL, type = "source")

Basic workflow

library(CRAFT)

set.seed(1)
series <- data.frame(
  asset_a = cumprod(1 + rnorm(100, 0.001, 0.01)),
  asset_b = cumprod(1 + rnorm(100, 0.0005, 0.012))
)

fit <- craft_fit(
  series,
  window = 5,
  n_draws = 100,
  n_factors = 1,
  min_segment = 5,
  max_regimes_per_factor = 3,
  n_testing = 0,
  verbose = FALSE
)

fit$valid_joint_acc
pred <- craft_predict(fit, n_draws = 25, seed = 1)

names(pred$return_dists)

Lower-level functions are available for custom pipelines:

• trajectory_embedding() builds aligned past and future trajectory matrices.
• svd_changepoint_regimes() detects regimes in transformed time-series data.
• regime_sampler() samples historical future-trajectory rows from regime probabilities.
• trajectory_forecast() fits smooth empirical forecast distributions.
• denoiser() selects informative numeric features before regime detection.
• time_volume_weights() creates volume-derived observation weights.

Hyperparameter tuning

craft_autotune() samples candidate settings and ranks their internal backtest scores. Lower scores are better; direction accuracy is negated for ranking.

tuned <- craft_autotune(
  series, n_models = 2, top_k = 1, n_testing = 2,
  search_space = list(transition_alpha = c(0.5, 1, 2)),
  fixed_args = list(window = 5, horizon = 5, n_draws = 40,
                   n_factors = 1, min_segment = 5,
                   max_regimes_per_factor = 3),
  seed = 123, verbose = FALSE
)
tuned$results
if (length(tuned$top_models)) {
  best_fit <- tuned$top_models[[1]]$fit
  prediction <- craft_predict(best_fit, n_draws = 25, seed = 7)
}

Search specifications may be candidate vectors, scalar constants, zero-argument sampler functions, or lists such as list(type = "numeric", min = 0.1, max = 10, log = TRUE). Sampled settings override fixed_args, which override arguments in .... An empty search_space = list() repeats the fixed settings with different model seeds. Optional search_parallel = TRUE uses two workers by default and disables nested backtest parallelism. Serial backtests are also the default during serial search; pass backtest_parallel = TRUE explicitly to enable them.

Keep horizon and backtest_min_train fixed when comparing different windows or segment lengths so candidates are evaluated on the same outcomes. Inspect tuned$failures before using the selected fit. These backtests select parameters; use a separate, later holdout period to assess the selected model’s performance.

License

GPL-3

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.