quickGAM: quickGAM

In shiggo/jezioro: Tools for Paleolimnological Analyses at the Paleoecological Environmental Assessment and Research Laboratory

Description

quickGAM is a function to fit a GAM and perform the related analyses described in Simpson (2018). This function essentially applies the code provided in that article's Supplementary Information, to a pair of vectors.

Usage

quickGAM(x, y, xLabel = NULL, yLabel = NULL, yRange = NULL, skipAuto = FALSE)

Arguments

x	vector containing the independent data/linear predictor
y	vector containing the dependent data/response variable
xLabel	(defaults to NULL): optional custom string to ensure consistent x-axis labels across the plots
yLabel	(defaults to NULL): optional custom string to ensure consistent y-axis labels across the plots
yRange	(defaults to NULL): optional custom range to ensure consistent y-axis across all plots
skipAuto	(Defaults to FALSE): logical to indicate whether to skip the autocorrelation calculation and plot

Details

Two input vectors are required (x and y, the independent and dependent variable respectively). The optional arguments allow customization of the x and y axes labels to ensure they remain consistent across all graphs. Similarly the y-range can be manually defined.

The function produces six plots:

• Basic scatterplot of the data

• Estimated CAR1 process from the GAM fitted to the data ( i.e. autocorrelation check)

• GAM-based trends fitted to the data

• Estimated trends (w 20 random draws of the GAM fit to the data)

• 95% simultaneous confidence intervals (light grey) and across-the-function confidence intervals (dark grey) on the estimated trends.

• Estimated first derivatives (black lines) and 95% simultaneous confidence intervals of the GAM trends fitted to the data. Where the simultaneous interval does not include 0, the models detect significant temporal change in the response.

Author(s)

Adam Jeziorski

References

Simpson GL (2018) Modelling Palaeoecological Time Series Using Generalised Additive Models. Frontiers in Ecology and Evolution [https://dx.doi.org/10.3389/fevo.2018.00149]

Examples

#Fit a GAM to the example chlorophyll a data in 'vrsChlaInput'
data(vrsChlaInput)
plotData <- vrsChla(vrsChlaInput)
#Values/labels to be preserved across all plots
indepVarLabel <- "Core Depth (cm)"
depenVarLabel <- expression(
 "VRS-Inferred Chlorophyll "~italic("a")~" (mg"%.%"g"^"-1"*textstyle(")"))
depenVarRange <- c(0, 0.0825)

#NOTE: drop=FALSE when reading in data columns preserves column names, allowing their use in plots.
quickGAM(plotData[,1, drop=FALSE],plotData[,2, drop=FALSE], xLabel = indepVarLabel, 
yLabel = depenVarLabel, yRange = depenVarRange)

shiggo/jezioro documentation built on Sept. 7, 2020, 7:34 p.m.