eqm_bc_5foldcv_simple: Calibrating the bias of observed data by empirical quantile...
In jieyu97/QCwind: Quality control procedures for citizen science wind observations

Performing empirical quantile distribution mapping to correct bias in the observation data, applying the 5-fold cross-validation; the "truth" quantiles are given for a reference.

1	eqm_bc_5foldcv_simple(data.obs, data.datetime, true.quantiles)

`data.obs`	a numeric vector of observation data needs to be corrected.
`data.datetime`	a sequence of timestamps in `data.obs`.
`true.quantiles`	a list of six numeric vectors of the "truth" quantiles to build the bias correction model, the corresponding cumulative probability stamps must be evenly spreaded between 0 and 1. If not satisfied, a pseudo "truth" observation numeric vector satisfying the quantiles should be provided. We suggest using quantiles with cumulative probability from 0.01 to 1 with interval length 0.01 (100 quantiles in total). We sort these quantiles to avoid crossovers.

a list of two parts: one is the output dataframe bc_output that includes both input and bias-corrected observation data; the other one is a large list bc.eqm_model contains all the empirical quantile mapping models in each turn of cross-validation during different season/period.

jieyu97/QCwind documentation built on June 18, 2021, 3:37 a.m.