select_harmonies: Select harmonies with significant patterns
In Sayani07/hakear:

Select harmonies with significant patterns

select_harmonies(
  .data,
  harmony_tbl = NULL,
  response = NULL,
  quantile_prob = seq(0.01, 0.99, 0.01),
  dist_ordered = TRUE,
  lambda = 0.67,
  nperm = 200,
  use_perm = TRUE,
  nsamp = 200
)

`.data`	a tsibble or data with already computed categories
`harmony_tbl`	A tibble containing one or more hamronies with facet_variable, x_variable, facet_levels and x_levels
`response`	the response variable
`quantile_prob`	numeric vector of probabilities with value #'in [0,1] whose sample quantiles are wanted. Default is set to #' "decile" plot
`dist_ordered`	if categories are ordered
`lambda`	value of tuning parameter for computing weighted pairwise distances
`nperm`	number of permutations for normalization
`use_perm`	should permutation approach for normalization be used
`nsamp`	number of permutation for computing the threshold

library(gravitas)
library(parallel)
library(dplyr)
library(tidyr)
sm <- smart_meter10 %>%
  filter(customer_id %in% c("10017994"))
harmonies <- sm %>%
  harmony(
    ugran = "year",
    filter_in = "wknd_wday",
    filter_out = c("hhour", "fortnight", "quarter", "semester")
  )
 harmonies1 <- harmonies %>% mutate(facet_variable = NA)
 h = harmonies1 %>% select(-facet_levels) %>% distinct() %>% mutate(facet_levels = NA)
all_harmony <- select_harmonies(sm,
  harmony_tbl = h,
  response = general_supply_kwh, nperm = 200, nsamp = 20
)
all_harmony2 <- select_harmonies(sm,
  harmony_tbl = harmonies,
  response = general_supply_kwh, nperm = 20, nsamp = 20
)