Skip to content

clipo/Seriation

Repository files navigation

Seriation

Algorithms, implementations, manuscripts, and test cases for iterative deterministic frequency seriation done by agglomeration.

These algorithms are intended to automate and extend the initial, mainly manual implementation of IDSS in Lipo (2001).

commandline:

python ./IDSS.py --inputfile=../testdata/pfg.txt --xyfile=../testdata/pfgXY.txt --largestonly=1 --mst=1 --screen=1

as a module:

import IDSS seriation= IDSS()

args={} args['inputfile'] ="../testdata/pfg.txt" args['xyfile']="../testdata/pfgXY.txt" args['largestonly']=1 args['screen']=1 args['mst']=1

seriation.seriate(args)

'''''

Directory Structure

  • python

Contains python modules for seriation. The perl script has been entirely rewritten into python and this is the version that we are using for most of the development.

  • analysis

Contains R, mathematica, and other scripts aimed at analyzing the basic problem and trial algorithms, before being implemented in software. Also will contain benchmarks and performance analysis of finished software versions, to document scaling of the algorithms and optimizations.

  • doc

Software documentation, installation instructions, FAQs, todo lists

  • java

Main root directory for a Java library implementing the seriation algorithms, for embedding in many types of systems. May also contain simple UI implementations (CLI, web app) to run seriations.

  • manuscripts

Directories containing journal manuscripts describing the algorithm or applying it to data. Each directory represents a separate paper and is self-contained.

  • output

Directory for writing finished seriation solutions for test data sets. This provides common configuration for software implementations and system tests.

  • perl

Perl implementation of the seriation algorithm. This serves as our initial testing ground and way of trying ideas for optimizations, and it will be useful long-term for small or "normal" data sets.

  • testdata

Repository of real and idealized data sets used to test the algorithm and implementations.

Authors

Carl P. Lipo, California State University at Long Beach Mark E. Madsen, University of Washington, Seattle

(1) Lipo, Carl P. 2001 Science, Style and the Study of Community Structure: An Example from the Central 1317 Mississippi River Valley. British Archaeological Reports, International Series, no. 918. Oxford.