Calculate the distance between two matrices.
Calculate the distance between two matrices.
Find the euclidean distance bewtween two matrices.
an adjacency matrix at step n
same adjacency matrix at step n+1
a normalized distance between m1 and m2
Use another object to speed up join between RDD.
Expand matrix
Expand matrix
an adjacency matrix
expanded adjacency matrix
Get epsilon coefficient
Get epsilon coefficient
Change an edge value to zero when the overall weight of this edge is less than a certain percentage
Get expansion rate
Get graph orientation strategy selected depending on graph nature
Get inflation rate
Get stop condition if MCL algorithm does not converge fairly quickly
Get weight of automatically added self loops in adjacency matrix rows
Available graph orientation strategy options.
Available graph orientation strategy options.
README.md for more details
Inflate matrix
Inflate matrix
Prune and normalization are applied locally (on each row). So we avoid two more complete scanning of adjacency matrix. As explained in issue #8, pruning is applied on expanded matrix, so we take advantage of natural normalized expansion state.
an adjacency matrix
inflated adjacency matrix
Normalize row
Normalize row
an unnormalized row of th adjacency matrix
normalized row
Normalize matrix
Normalize matrix
an unnormalized adjacency matrix
normalized adjacency matrix
Remove weakest connections from a row
Remove weakest connections from a row
Connections weight in adjacency matrix which is inferior to a very small value is set to 0
a row of the adjacency matrix
sparsed row
Add more complex pruning strategies.
http://micans.org/mcl/index.html
Train MCL algorithm.
Train MCL algorithm.
a graph to partitioned
an MCLModel where each node is associated to one or more clusters
Set the minimum percentage to get an edge weight to zero.
Set the minimum percentage to get an edge weight to zero.
Default: 0.01.
epsilon must be higher than 0 and lower than 1
Set the expansion rate.
Set the expansion rate.
Default: 2.
expansionRate must be higher than 1
Set graph orientation strategy.
Set graph orientation strategy.
Default: undirected.
graphOrientationStrategy must be contained in graphOrientationStrategyOption
Set the inflation rate.
Set the inflation rate.
Default: 2.
inflationRate must be higher than 0
Set maximum number of iterations.
Set maximum number of iterations.
Default: 10.
maxIterations must be higher than 0
Set self loops weights.
Set self loops weights.
Default: 0.1.
selfLoopWeight must be higher than 0 and lower than 1
A clustering model for MCL.
README.md for more details on theory