Add time series features calculated by the tsfeatures package
Source:R/add-tsfeatures.R
add_tsfeatures.Rd
add_tsfeatures()
adds a set of calculated features from the tsfeatures package for each time series in the group. These features provide information about various characteristics of the time series.
Arguments
- .tbl
tibble; data with date (class: Date), value (class: numeric), and group (class: character) columns
- ...
character; one or more unquoted grouping columns
- standardize
logical; If TRUE (default), the function with standardize each feature.
- parallel
logical; If TRUE, features will be calculated in parallel. Default is FALSE.
Details
Function can be used with a global forecasting method or for EDA. See the tsfeatures website for more details on these features.
References
Pablo Montero-Manso, Rob J. Hyndman, Principles and algorithms for forecasting groups of time series: Locality and globality, International Journal of Forecasting, 2021 link
Examples
library(dplyr, warn.conflicts = FALSE)
group_ts_tbl <- tsbox::ts_tbl(fpp2::arrivals)
head(group_ts_tbl)
#> # A tibble: 6 × 3
#> id time value
#> <chr> <date> <dbl>
#> 1 Japan 1981-01-01 14.8
#> 2 Japan 1981-04-01 9.32
#> 3 Japan 1981-07-01 10.2
#> 4 Japan 1981-10-01 19.5
#> 5 Japan 1982-01-01 17.1
#> 6 Japan 1982-04-01 10.6
new_tbl <- add_tsfeatures(group_ts_tbl, id)
head(new_tbl)
#> # A tibble: 6 × 23
#> id time value frequency nperiods seasonal_period trend spike
#> <chr> <date> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 Japan 1981-01-01 14.8 4 1 4 0.327 0.0853
#> 2 Japan 1981-04-01 9.32 4 1 4 0.327 0.0853
#> 3 Japan 1981-07-01 10.2 4 1 4 0.327 0.0853
#> 4 Japan 1981-10-01 19.5 4 1 4 0.327 0.0853
#> 5 Japan 1982-01-01 17.1 4 1 4 0.327 0.0853
#> 6 Japan 1982-04-01 10.6 4 1 4 0.327 0.0853
#> # ℹ 15 more variables: linearity <dbl>, curvature <dbl>, e_acf1 <dbl>,
#> # e_acf10 <dbl>, seasonal_strength <dbl>, peak <dbl>, trough <dbl>,
#> # entropy <dbl>, x_acf1 <dbl>, x_acf10 <dbl>, diff1_acf1 <dbl>,
#> # diff1_acf10 <dbl>, diff2_acf1 <dbl>, diff2_acf10 <dbl>, seas_acf1 <dbl>