hankruiger/tsnet

Source code accompanying the submission to EuroVis 2017

tsNET

Graph Layouts by t-SNE

usage: tsnet.py [-h] [--star] [--perplexity PERPLEXITY]
                [--learning_rate LEARNING_RATE] [--output OUTPUT]
                input_graph

Read a graph, and produce a layout with tsNET(*).

positional arguments:
  input_graph

optional arguments:
  -h, --help            show this help message and exit
  --star                Use the tsNET* scheme. (Requires PivotMDS layout in
                        ./pivotmds_layouts/ as initialization.) Note: Use
                        higher learning rates for larger graphs, for faster
                        convergence.
  --perplexity PERPLEXITY, -p PERPLEXITY
                        Perplexity parameter.
  --learning_rate LEARNING_RATE, -l LEARNING_RATE
                        Learning rate (hyper)parameter for optimization.
  --output OUTPUT, -o OUTPUT
                        Save layout to the specified file.

Example:

# Read the input graph dwt_72, and save the output in ./output.vna
./tsnet.py graphs/dwt_72.vna --output ./output.vna

Dependencies

• python3
• numpy
• matplotlib
• graph-tool
• theano
• scikit-learn

Example