Matrix Exploratory Data Analysis
Switch branches/tags
Clone or download
Pull request Compare This branch is 117 commits ahead, 1 commit behind MrAE:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
R
docker
examples
inst/extdata
man
python
tests
.Rbuildignore
.gitignore
DESCRIPTION
LICENSE
NAMESPACE
README.md

README.md

README

meda: Matrix Exploratory Data Analysis is an R package for exploring data in the form of matrices.

for the python version, go here https://github.com/neurodata-nomads/pymeda

INSTALLATION

Currently there seems to be a problem installing in RStudio. The best way seems to be installing from R via command line.

To download and install run the following in R:

require(devtools)
devtools::install_github("neurodata/meda")

Run on your own data

To run meda on your own data do the following. Supposing your data is in a csv file with a header:

require(meda)
dat <- read.csv("fileName.csv", header=TRUE)

Example using Fisher's Iris data

First we center the feature columns and then we proceed with exploring the data.

dato <- iris[, -5]
dat <- data.frame(scale(dato, center = TRUE, scale = FALSE))
truth <- iris[, 5]

## optional column classes or feature classes
ccol <- rep(c("darkgreen", "purple"), 2)

plot(mlocation(dato, ccol = ccol))
plot(d1heat(dat, ccol = ccol))
plot(outliers(dat))
plot(medacor(dat, ccol = ccol))
plot(cumvar(dat))
pairhex(dat)

h <- hmc(dat, ccol = ccol)
plot(h, pch = truth)
plotDend(h)
stackM(h, centered = TRUE, ccol = ccol, depth = h$dat$height)
stackM(h, centered = TRUE, ccol = ccol)
clusterMeans(h)

Given a new set of n samples in R^d

follow instructions here