Skip to content

Merging code from Dev wordsi into master#19

Open
fozziethebeat wants to merge 142 commits into
masterfrom
dev-wordsi
Open

Merging code from Dev wordsi into master#19
fozziethebeat wants to merge 142 commits into
masterfrom
dev-wordsi

Conversation

@fozziethebeat

Copy link
Copy Markdown
Owner

Dev-Wordsi contains a multitude of small fixes and a bulk of new code centered around consensus clustering. This also makes a significant change to the Assignments interface to make it easier and less cumbersome to use.

The new Consensus Clustering code relies on a new object called a partition, which is an easy to use representation of a clustering solution that includes some addiition helper methods not useful to general clustering Assignments. The Consensus merging algorithms use a comparison function between partitions and a combining algorithm to generate an improved Partition.

In addition to this major update, there are smaller bug fixes and minor refactorings needed within the consensus clustering code.

…ltered bigram matrix, Updating similarity to compile, Fixing some minor bugs in consensus clustering
…javadoc. Also update some scala code with comments.

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I realize this might be a pain, but since we have a graph package, could we somehow retrofit this to use a Graph instance and be in the graph package?

Alternatively, we could wrap a Matrix as a Graph and implement it that way?

Copy link
Copy Markdown
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'd rather not change this to use the graph interface for two reasons:

  • It currently implements the Clustering interface, which is really nice and portable
  • This provides an alternative implementation with slightly different guarantees as far as internal data structure

If we want cross compatibility, we could easily write wrappers between graphs and matrices, but that should be in a separate change.

@davidjurgens

Copy link
Copy Markdown
Collaborator

Overall, the changes look good. I have a few concerns where the graph package and some of the matrix-as-graph classes overlap in functionality. It would be good to present a common theme for those.

@fozziethebeat

Copy link
Copy Markdown
Owner Author

I still need to go through and fix the javadocs and the minor suggestions, but I put in my thoughts on some of the larger suggestions.

@fozziethebeat

Copy link
Copy Markdown
Owner Author

Most of the javadoc and simple cleaning changes fixed. There's still a few more to be handled.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants