dbl_block_bs: Double seasonal block bootstrap
In camroach87/gefcom2017: Functions for GEFCOM 2017

Description Usage Arguments Details Value Author(s)

Generates bootstrapped dates using a double seasonal block bootstrap approach.

1 2	dbl_block_bs(dates, start_date = today(), end_date = today() + 30, n_sims = 100, avg_block_len = 14, delta_loc = 3, delta_len = 3)

`start_date`	start date for period we want bootstrap samples for.
`end_date`	end date for period we want bootstrap samples for.
`n_sims`	number of simulations. Defaults to 100 simulations.
`avg_block_len`	average length of blocks. Defaults to 14 days.
`delta_loc`	amount to randomise current location by when sampling blocks. See details.
`delta_len`	amount to randomise block length by. See details.
`date_series`	array of dates.

delta_loc controls how much the current location in the simulated year should be shifted by when sampling from historical data. The shifting amount is seleted from a discrete Uniform(-delta_loc, delta_loc) distribution. Similarly, delta_len controls how much the length of the bootstrap block should be. A smapled block will have length equal to avg_block_len + Uniform(-delta_len, delta_len).

A data frame with two columns. One for simulation number and another for bootstrapped dates.

Cameron Roach

camroach87/gefcom2017 documentation built on March 21, 2020, 11:46 p.m.