Please use MdSlicing.jl now.

Multi-dimensional Slicing

A library for computing 1D and 2D slices through multi-dimensional datasets.

Installation

Installation is via devtools

install_github("gabysbrain/multid-slicing")

1D slice example

1. If your data is a set of samples in a table, rather than a function then build a regression model on your dataset so you have a function to visualize. Divide your data into X which is a table of all independent variables and Y which is a vector of a single dependent variable. For this example, the regression is done using a Gaussian process model via the mlegp library.

   library(mlegp)
   
   m = mlegp(X, Y)
   f = function(x) {predict(m, x)}
   

2. You will also need to create a problemSpec instance to keep track of the names and limits of each dimension.

   library(multidslicing)
   lims = createProblemSpec(x1=c(-1, 1), x2=c(-1, 1), x3=c(-1, 1))
   

3. Create a set of slices from the function you have. Here, we create 50 slices of the function for each dimension.

   slices = sliceplorer(f, lims, 50)
   

4. Now plot the slices

   plot(slices)
   

License

This project is licensed under a BSD License --- see the LICENSE.md file for details.