LDAcrop.plus: Linear discriminant analysis based on attributes of weed...

In elizabethastroud/Cropprocessing: Data organisation, classification and visualisation of archaeobotanical data to understand crop processing stage

Linear discriminant analysis based on attributes of weed seeds to classify taphonomic pathway (crop processing vs other routes)

Description

This function conducts linear discriminant analysis using ethnographic crop processing data of weed seeds attributes. This function is a modification of LDAcrop.pro, and uses the entered archaeobotanical data as well as the ethnographic data during the discrimination stage to create a model. The entered archaeobotanical data is then reclassified against that model, allowing the archaeobotanical samples to be classified as 1 of five groups: archaeological, winnowing by-product, coarse sieve by-product, fine sieve by-product, fine sieve product. The function provides the classification, posterior probabilities of such classifications, and the discriminant score of the entered samples.

Usage

LDAcrop.plus(x)

Arguments

x	The archaeobotanical dataset

Details

The archaeobotanical dataset needs to have been transformed and organised with columns labelled and in the order of: BHH,BFH,SHH,SHL,SFH,SFL.The first column of the dataframe should be the sample names. Transformation can be done manually following (insert reference) or through the use of crop.dataorg which can transform a raw archaeobotanical dataset.

Value

Results table: (note the * asterisked columns appear in console output and are used for interpretation, and graphing. Non-asterisked columns provide additional details regarding standardised and unstandardised results

Samples	the archaeobotanical sample names from the first column of the entered dataset (x)
Class_std*	The standardised classification of the samples as either 1, 2, 3, or 4. 1= winnowing by-product, 2= coarse sieve by-product, 3= fine sieve by-product and 4= fine sieve product
Prob.1_std*	the standardised posterior probability of the sample being classified as group 1
Prob.2_std*	the standardised posterior probability of the sample being classified as group 2
Prob.3_std*	the standardised posterior probability of the sample being classified as group 3
Prob.4_std*	the standardised posterior probability of the sample being classified as group 4
ld1_std	the standardised linear discriminant score for function 1
ld2_std	the standardised linear discriminant score for function 2
ld3_std	the standardised linear discriminant score for function 3
Class	the unstandardised classification of the samples
Prob.1	the unstandardised posterior probability of the sample being classified as group 1
Prob.2	the unstandardised posterior probability of the sample being classified as group 2
Prob.3	the unstandardised posterior probability of the sample being classified as group 3
Prob.4	the unstandardised posterior probability of the sample being classified as group 4
LD1*	the unstandardised linear discriminant score for function 1
LD2*	the unstandardised linear discriminant score for function 2
LD3*	the unstandardised linear discriminant score for function 3

Classification table: showing the count and percentage of samples classified as one of four crop processing groups - as shown in the Class_std column

winnowing by-products	the count and percentage of samples classified as group 1
Coarse sieve by-product	the count and percentage of samples classified as group 2
Fine sieve by-product	the count and percentage of samples classified as group 3
Fine sieve product	the count and percentage of samples classified as group 4

Author(s)

Elizabeth Stroud

References

Charles, M., 1998. Fodder from dung: the recognition and interpretation of dung-derived plant material from archaeological sites, Environmental Archaeology, 1:1, 111-122

See Also

LDAcrop.pro, crop.dataorg

Examples

## Create random dataset for example

BHH<-runif(20, min=0, max=7)
BFH<-runif(20, min=0, max=24)
SHH<-runif(20, min=1, max=13)
SHL<-runif(20, min=0.5, max=17)
SFH<-runif(20, min=1, max=22)
SFL<-runif(20, min=1, max=8)
Samples<-c("s1","s2","s3","s4","s5","s6","s7","s8","s9","s10","s11","s12","s13",
"s14","s15","s16","s17","s18","s19","s20")
data<-data.frame(Samples,BHH,BFH,SHH,SHL,SFH,SFL)

## function usage

LDAcrop.plus(data)

elizabethastroud/Cropprocessing documentation built on Sept. 27, 2024, 3:03 p.m.