README.md
In bwrc/semigeom-r: Semigeometric tiling of data matrices

semigeom

The semigeom package provides functions for performing semigeometric tiling of data matrices. The basic idea is to find regions in data matrices where the density of some feature, e.g., ones in a binary matrix, is large enough, as specified by some threshold. A brief illustrative example is provided below. The exact technique is described in this paper:

Henelius, Andreas, Isak Karlsson, Panagiotis Papapetrou, Antti Ukkonen, and Kai Puolamäki. Semigeometric Tiling of Event Sequences. ECML PKDD 2016, Part I, LNAI 9851, pp. 329-344, 2016.

To get a BibTex entry in R type citation("semigeom") when the package is installed.

The development version of the semigeom package can be installed from GitHub as follows.

First install the devtools-package and load it:

install.packages("devtools")
library(devtools)

You can now install the semigeom package:

install_github("bwrc/semigeom-r")

This is a short example demonstrating use of the library.

## Load the library
library(semigeom)

## Create some example data
data <- make_dataset_list_synthetic()
X <- data$synthetic1

## Perform the tiling
tiles_mt <- max_tile(X, K = 4, alpha = 0.75)
tiles_gt <- global_tile(X, K = 5, alpha = 0.75)

## Plot the tiles
par(mfcol = c(1, 2))
plot_binary_matrix(X, tiles_mt, main = "MaxTile")
plot_binary_matrix(X, tiles_gt, main = "GlobalTile")

This produces the figure Tiles