Skip to content
/ DA Public

DA incorporates the commonly used linear and non-linear, local and global supervised learning approaches (discriminant analysis). These discriminant analyses can be used to do ecological and evolutionary inference. We show the examples of demographic history inference, species identification, and population structure inference in the vignettes …

License

Notifications You must be signed in to change notification settings

xinghuq/DA

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

64 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DA: Discriminant analysis

For a more detailed tutorial on population structure and community structure inference please visit: https://xinghuq.github.io/DA/

DAPC, DAKPC, LFDA, LFDAKPC, KLFDA

This package implements five supervised learning approaches that are suitable for ecological and evolutionary inference both visually and statistically. Five approaches in our vignette: Linear discriminant analysis of principal components (DAPC), linear discriminant analysis of kernel principal components (DAKPC), local (Fisher) linear discriminant analysis (LFDA), local (Fisher) discriminant analysis of kernel principal components (LFDAKPC) and kernel local (Fisher) linear discriminant analysis (KLFDA).

Welcome any feedback and pull request.

Install the package from github:

library(devtools)

install_github("xinghuq/DA")

library("DA")

Basic examples using the irish data

x <- iris[,-5] # this matrix contains all the predictors to be transformed
y <- iris[,5] # this should be a vector that represents different classes

Discriminant analysis of principal components (DAPC)

library(adegenet)
iris_dapc=dapc(iris[,-5],grp=iris[,5],n.pca=3, n.da=3)

#plot the data projection on the components
library(plotly)
   cols=rainbow(length(unique(iris[,5])))
   p1 <- plot_ly(as.data.frame(iris_dapc$ind.coord), x =iris_dapc$ind.coord[,1], y =iris_dapc$ind.coord[,2], color = iris[,5],colors=cols[iris[,5]],symbol = iris[,5],symbols = 1:3L) %>% 
     add_markers() %>%
     layout(scene = list(xaxis = list(title = 'LDA1'),
                         yaxis = list(title = 'LDA2')))
p1

Discriminant analysis of kernel principal components (DAKPC)

iris_ldakpc=LDAKPC(iris[,-5],y=iris[,5],n.pc=3)

 p2 <- plot_ly(as.data.frame(iris_ldakpc$LDs), x =iris_ldakpc$LDs[,1], y =iris_ldakpc$LDs[,2], color = iris[,5],colors=cols[iris[,5]],symbol = iris[,5],symbols = 1:3L) %>% 
     add_markers() %>%
     layout(scene = list(xaxis = list(title = 'LDA1'),
                         yaxis = list(title = 'LDA2')))
p2

Local Fisher Discriminant Analysis(LFDA)

iris_lfda=LFDA(iris[,-5],y=iris[,5],r=3,tol=1)

 p3 <- plot_ly(as.data.frame(iris_lfda$Z), x =iris_lfda$Z[,1], y =iris_lfda$Z[,2], color = iris[,5],colors=cols[iris[,5]],symbol = iris[,5],symbols = 1:3L) %>% 
     add_markers() %>%
     layout(scene = list(xaxis = list(title = 'LDA1'),
                         yaxis = list(title = 'LDA2')))
p3

Local (fisher) discriminant analysis of kernel principal components (LFDAKPC)

iris_lfdakpc=LFDAKPC(iris[,-5],y=iris[,5],n.pc=3,tol=1)

 p4 <- plot_ly(as.data.frame(iris_lfdakpc$$LFDAKPC$Z), x =iris_lfdakpc$$LFDAKPC$Z[,1], y =iris_lfdakpc$$LFDAKPC$Z[,2], color = iris[,5],colors=cols[iris[,5]],symbol = iris[,5],symbols = 1:3L) %>% 
     add_markers() %>%
     layout(scene = list(xaxis = list(title = 'LDA1'),
                         yaxis = list(title = 'LDA2')))
p4

Kernel Local Fisher Discriminant Analysis(KLFDA)

The default kernel is polydot(degree = 1, scale = 1, offset = 1). Users can set the kernel based on their own purpose.

iris_klfda=KLFDA(as.matrix(iris[,-5]),as.matrix(iris[,5]),r=3,tol=1E-10,prior = NULL)


p5 <- plot_ly(as.data.frame(iris_klfda$Z), x =iris_klfda$Z[,1], y =iris_klfda$Z[,2], color = iris[,5],colors=cols[iris[,5]],symbol = iris[,5],symbols = 1:3L) %>% 
     add_markers() %>%
     layout(scene = list(xaxis = list(title = 'LDA1'),
                         yaxis = list(title = 'LDA2')))
p5

Note we did not show the discrimination results,users can look into the discriminated classes and their posterior possibility from the results.

Citation

Qin. X, Wu. M, Lock, T. R, Kallenbach, R. 2020. DA: Ecological and evolutionary inference using machine learning approaches both visually and statistically. R package version 1.2.0.

About

DA incorporates the commonly used linear and non-linear, local and global supervised learning approaches (discriminant analysis). These discriminant analyses can be used to do ecological and evolutionary inference. We show the examples of demographic history inference, species identification, and population structure inference in the vignettes …

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages