mds: Multidimensional scaling

In pdi: Phenotypic Index Measures for Oak Decline Severity

Description

perform multidimensional scaling of random forest proximities

Usage

mds(rfModels, dimensions = 2)

Arguments

rfModels	list containing random forest models as returned by rf()
dimensions	number of dimensions to scale to

Examples

library(dplyr)

## Retrieve file paths for example data
files <- list.files(system.file('phenotypeDataCollectionSheets',
  package = 'pdi'),full.names = TRUE)

## Prepare data
d <- map(files,readPhenotypeSheet) %>%
  map(preparePhenotypeData) %>%
  bind_rows() %>%
  siteAdjustment() %>%
   mutate(`Live crown ratio (%)` = liveCrownRatio(`Total height (m)`,
     `Lower crown height (m)`),
     `Crown condition (%)` = crownCondition(`Missing crown (%)`,
                               `Crown transparency (%)`),
     `Crown volume (m^3)` = crownVolume(`Crown radius (m)`,
                               `Total height (m)`,
                               `Lower crown height (m)`,
                               `Crown condition (%)`),
     `Bleed prevalence (%)` = bleedPrevalence(`Active bleed length (mm)`,
                               `Active bleeds`,
                               `Black staining length (mm)`,
                               `Black staining`,
                               `Diameter at breast height (m)`),
     `Agrilus exit hole density (m^-2)` = agrilusExitHoleDensity(`Agrilus exit holes`,
                               `Diameter at breast height (m)`)
)

t <- makeAnalysisTable(d)

## Generate random forest models
m <- rf(t,cls = NULL,nreps = 10)

mds_data <- mds(m,2)

pdi documentation built on Feb. 10, 2021, 1:08 a.m.