autocorr

Test for Autocorrelation

Description

Tests the final.csv created with ‘fin.csv’ on autocorrelation to produce reliable models.

Usage

autocorr(
  method = "monthly",
  resp,
  pred,
  plot.corrplot = TRUE,
  corrplot = "Coef"
)

Arguments

method character. Choose the time scale your data is preserved in. Either “annual”, “monthly” or “daily”.
resp numerical. Vector or single input of the columns in the final.csv that contain your sensor data (“response variables”). The function will create one file per variable.
pred numerical. Vector or single input. The columns of your predictor variables, that you want to test for autocorrelation with the response variables.
plot.corrplot logical. Should correlation matrices be plotted?
corrplot character. Vector or single input. If plot.corrplot is true, you can choose the design of the correlation plot. You can choose from “coef”, “crossout”, “blank”. Default is “coef”.

Value

One .csv file per response variable. These will later be used when ‘autocorrelation’ is set ‘TRUE’ during ‘calc.model’.

Examples

## Not run: 
# Test data for autocorrelation after running fin.csv
autocorr(method = "monthly",
         resp = 5,
         pred = c(8:24),
         plot.corrplot = FALSE)

## End(Not run)