Help required with understanding path-degree product calculation

[yuvaramsingh94]

Hi,
This is Yuvaram a Master's student working on heterogeneous network. I am trying to understand the Path-degree product calculation from your Paper. I don't understand what is metaedge specific degrees along the path and why each edge contributes to 2. Can you provide an example or a simple calculation. This will help me in using your work for my project.

Thanks
Yuvaram

[dhimmel]

Figure 2D from the paper is the best visualization:

Since each edge has a source and target node, two degrees exist for each edge. These degrees are metaedge specific (only counting edges of the same type).

Here are two implementations from this library if this helps understand. First the networkx/python implementation:

hetnetpy/hetnetpy/pathtools.py

Lines 32 to 34 in e1ff1f8

	for edge in path:
	source_edges = edge.source.get_edges(edge.metaedge, exclude_masked)
	target_edges = edge.target.get_edges(edge.metaedge.inverse, exclude_masked)

And second the neo4j/cypher implementation

hetnetpy/test/pathtools_test.py

Lines 82 to 92 in e1ff1f8

	WITH
	[
	size((n0)-[:BINDS_CbG]-()),
	size(()-[:BINDS_CbG]-(n1)),
	size((n1)-[:PARTICIPATES_GpPW]-()),
	size(()-[:PARTICIPATES_GpPW]-(n2)),
	size((n2)-[:PARTICIPATES_GpPW]-()),
	size(()-[:PARTICIPATES_GpPW]-(n3)),
	size((n3)-[:ASSOCIATES_DaG]-()),
	size(()-[:ASSOCIATES_DaG]-(n4))
	] AS degrees, path

[yuvaramsingh94]

Thanks for quick reply any clearing the doubt.