Wish List

page VFML

The windows distribution needs to be brought up to date.

File batchtest

I think the input format for batchtest is a little brittle and it could use some improvement

File BeliefNet.h

A version of this that uses DecisionTree local models instead of full CPTs. This would also need a new structure learning tool (a modification of bnlearn).

File beliefnetcorrupt

This tool does not do anything smart with parameters when it changes the structure, and it should.

File beliefnetscore

Move the comparision mode from this tool into a new tool, beliefnetcompare, and have that tool do more interesting comparisions (e.g. measure the KL-distance between the distriutions encoded in the networks).

File bindata

that this tool would have more methods for selecting bin boundaries, for example to reduce entropy.

File bnlearn

A version of this program that is intelligent about dealing with unobserved (or partially observed) variables.

File cvfdt

Modify this learner to work with continuous attributes.

An API to this learner like the one to learning BeliefNet structure in beliefnet-engine.h

File DecisionTree.h

A standard in memory decision tree induction algorithm. Maybe the best starting point would be the decisionstump learner.

This isn't the right place for this wish, but it would be nice to have a RuleSet structure similar to this DecisionTree structure

File HashTable.h

A cleaner hash table interface (without the compare function). I think that the sprint learner is the only thing that uses this currently so there isn't much to change to fix this.

File kmeans

Modify this learner to work with discrete attributes.

Move the features from vfkm into this learner because this learner will be much easier to modify than that one for new users.

File naivebayes

Modify this learner to work with continuous attributes.

File REPrune.h

More pruning methods, and more sophisticated REPruning (for example, all the leaves to prune could be selected in a single pass over the tree).

File shuffledata

for a version of this tool that works on disk and does not need to load data into RAM.

File stats.h

The StatTracker would track more interesting things.

File vfbn1

An API to this learner like the one to learning BeliefNet structure in beliefnet-engine.h

File vfbn2

An API to this learner like the one to learning BeliefNet structure in beliefnet-engine.h

File vfdt-engine.h

A function that checkpoints the learning procedure to disk so that it can be restored at a later time. I think the hard part of this would be checkpointing the ExampleGroupStats structure.

File vfem

Modify this learner to also learn the variances of the Gaussians.

---

Generated for VFML by hosted by