| Title: | 'MiniRocket': A Very Fast (Almost) Deterministic Transform for Time Series Classification |
|---|---|
| Description: | High-performance R and C++ implementation using 'OpenMP' parallelization for the 'MiniRocket' algorithm. Extracts features from univariate time series for downstream classification as described in 'Dempster et al.' (2021) (<doi:10.1145/3447548.3467231>). |
| Authors: | Manuel Villacorta Tilve [aut, cre] |
| Maintainer: | Manuel Villacorta Tilve <[email protected]> |
| License: | GPL (>= 3) |
| Version: | 0.1.0 |
| Built: | 2026-10-08 10:58:15 UTC |
| Source: | https://github.com/cran/minirocketR |
Fit a MiniRocket Feature Transformer
minirocket_fit( X, num_features = 10000, max_dilations_per_kernel = 32, seed = NULL )minirocket_fit( X, num_features = 10000, max_dilations_per_kernel = 32, seed = NULL )
X |
A numeric matrix of dimensions N x L. |
num_features |
Integer. Target number of features (default: 10000, yields 9996). |
max_dilations_per_kernel |
Integer. Maximum number of dilations considered per kernel (default: 32, as in the reference MiniRocket implementation). |
seed |
Optional integer random seed. |
An object of class minirocket.
# Create a synthetic matrix of 10 time series of length 50. set.seed(42) X <- matrix(rnorm(10 * 50), nrow = 10, ncol = 50) # Fit the MiniRocket model. mod <- minirocket_fit(X, num_features = 1000, seed = 42) print(mod)# Create a synthetic matrix of 10 time series of length 50. set.seed(42) X <- matrix(rnorm(10 * 50), nrow = 10, ncol = 50) # Fit the MiniRocket model. mod <- minirocket_fit(X, num_features = 1000, seed = 42) print(mod)
Applies a fitted minirocket model to extract PPV feature matrices
using OpenMP parallelization in C++.
minirocket_transform(object, X, num_threads = 1, ...)minirocket_transform(object, X, num_threads = 1, ...)
object |
A fitted |
X |
A numeric matrix of dimensions N x L. |
num_threads |
Integer specifying CPU threads (default 1). |
... |
Unused additional arguments. |
A numeric matrix of size N x 9996 containing PPV features (by default).
set.seed(42) X <- matrix(rnorm(10 * 50), nrow = 10, ncol = 50) # Training and transform. modelo <- minirocket_fit(X, num_features = 1000, seed = 42) X_feat <- minirocket_transform(modelo, X) dim(X_feat)set.seed(42) X <- matrix(rnorm(10 * 50), nrow = 10, ncol = 50) # Training and transform. modelo <- minirocket_fit(X, num_features = 1000, seed = 42) X_feat <- minirocket_transform(modelo, X) dim(X_feat)