powdR (version 1.3.0)

xrpd_pca: PCA of XRPD data

Description

xrpd_pca is used to apply principal component analysis to X-ray powder diffraction data.

Usage

xrpd_pca(x, mean_center, bin_size, root_transform, components)

Arguments

x

A multiXY list containing the XRPD data, where each item in the list is a 2 column XY dataframe defining the x (2theta) and y (counts) axes of each measurement. Each item in the list must have a name corresponding to a unique sample ID.

mean_center

A logical argument defining whether mean centering is applied to the XRPD data (default = TRUE).

bin_size

An integer between 1 and 10 defining whether to bin the XRPD data to a lower resolution. This bin_size defines the number of data points used in each bin.

root_transform

An integer between 1 and 8 defining the root transform to apply to the XRPD data

components

An integer defining the number of principal components to include in the output. Must be at least 1 less than the number of XRPD patterns in the dataset (the default).

Value

a list with components:

coords

a dataframe containing the sample ID's for each sample and the PCA coordinates for each dimension

loadings

a dataframe containing the 2theta axis and the loading of each dimension

eig

a dataframe summarising the variance explained by each dimension

Details

Applies data pre-treatment and principal components analysis to XRPD data based based on the protocols detailed in Butler et al. (2020).

References

Butler, B.M., Sila, A.M., Shepherd, K.D., Nyambura, M., Gilmore, C.J., Kourkoumelis, N., Hillier, S., 2019. Pre-treatment of soil X-ray powder diffraction data for cluster analysis. Geoderma 337, 413-424. doi:10.4236/ampc.2013.31A007

Examples

# NOT RUN {
data(rockjock_mixtures)

x1 <- xrpd_pca(rockjock_mixtures,
               mean_center = TRUE,
               bin_size = 1,
               root_transform = 1)

#Plot the loading of dimension 1

plot(x = x1$loadings$tth,
     y = x1$loadings$Dim.1,
     type = "l")

# }
# NOT RUN {
#Fit loading 1 to the rockjock library
f1 <- fps_lm(rockjock,
             smpl = data.frame("tth" = x1$loadings$tth,
                               "counts" = x1$loadings$Dim.1),
             refs = rockjock$phases$phase_id,
             std = "QUARTZ",
             align = 0,
             p = 0.05)

plot(f1, wavelength = "Cu", interactive = TRUE)
# }
# NOT RUN {
# }