https://github.com/albertmeronyo/somoclu-dyldo

https://github.com/albertmeronyo/somoclu-dyldo

Last synced: 10 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/albertmeronyo/somoclu-dyldo
• Owner: albertmeronyo
• License: mit
• Created: 2016-07-28T09:58:07.000Z (over 8 years ago)
• Default Branch: master
• Last Pushed: 2016-10-25T10:25:48.000Z (about 8 years ago)
• Last Synced: 2024-10-05T23:21:30.291Z (3 months ago)
• Language: Python
• Size: 9.57 MB
• Stars: 0
• Watchers: 2
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.md

Awesome Lists containing this project

README

        
# somoclu-dyldo

This repo contains data on some number-crunching performed on 79 snapshots of the [DyLDO dataset](http://swse.deri.org/dyldo/) to be processed by the ESOM tool [somoclu](https://github.com/peterwittek/somoclu).

The root directory contains the scripts:

- `order.py`: orders indices of sparse-matrix files

- `randbook.py`: generates a random codebook for the first somoclu map

- `somoclu.py`: runs somoclu on a set of pre-computed sparse matrices (one per DyLDO snapshot)

- `train.py`: likewise, considering the first snapshot and randomly generated codebook

- `uniqs.py`: generates lists of unique graph names and predicates, and computes the sparse matrices

- `viz.py`: generates image files for the processed maps using matplotlib/numpy (these can be pasted using e.g. `convert -delay 30 '*.png' movie.avi`)

And the data files:

- `uniqGRAPHS`: sorted unique graph names in all 79 snapshots (intersection)

- `uniqPREDS`: sorted unique predicate URIs in all 79 snapshots (intersection)

The `data` directory contains the generated sparse matrices.