quantile_gran: compute distances and groups from algorithm based on raw...

In Sayani07/gracsr:

Description

compute distances and groups from algorithm based on raw distributions

Usage

quantile_gran(
  .data,
  gran1 = NULL,
  gran2 = NULL,
  response = NULL,
  quantile_prob_val = seq(0.1, 0.9, 0.1),
  group = NULL
)

Arguments

.data	a tsibble
gran1	one granularity e.g. hour_day, day_week, wknd_wday
gran2	one granularity distinct from gran1
response	measured variable
quantile_prob_val	values of probability for which distances between quantiles would be computed
group	NULL if quantiles to be obtained for the key variable and the column name of the group variable if quantiles to be obtained for the group.

Examples

library(gravitas)
library(tidyverse)
sm <- smart_meter10 %>%
filter(customer_id %in% c("10006704", "10017936","10006414", "10018250"))
gran1 = "hour_day"
gran2 = NULL
response = "general_supply_kwh"
dist_gran(sm, "hour_day")
dist_gran(sm, "month_year")
sm %>% quantile_gran(gran1 = "hour_day")
group = tibble(customer_id = c("10006704", "10017936", "10006414", "10018250"), group = c(1,2,1,1))
sm_group <- sm %>% left_join(group)
.data <- sm
quantile_gran(sm_group, gran1, group = "group") # obtain quantiles for group
quantile_gran(sm, gran1, group = NULL) # obtain quantiles for customer

Sayani07/gracsr documentation built on Dec. 18, 2021, 12:59 p.m.