Aleph is both a multi-platform machine learning framework aimed at simplicity and performance, and a library of selected state-of-the-art algorithms.
semi-supervised algorithms: graph label propagation, discrete regularization, etc.
large-scale linear algorithms: logistic linear regression, stochastic gradient descent linear SVM, etc.
wrappers to well-known tools: libsvm, SVMlight, etc.
graph-based algorithms: random walks, absorbing random walks, etc.
feature selection statistics: infogain, cross entropy, chi-squared, etc.
convenience validation utilities: several splitting methods, several scoring functions
fast vector and matrix implementations: based on matrix toolkits for java, but with a few optimizations on top of it
fast on-the-fly operations over datasets, instances and features: based on the concept of views over those first-class objects in the framework
Most importantly, the framework features a clean design and is therefore easily extensible.
Aleph 0.6 is faster, more stable and better documented than the previous version.
- Changes to previous version:
Initial Announcement on mloss.org.
No one has posted any comments yet. Perhaps you'd like to be the first?
Leave a comment
You must be logged in to post comments.