About: The Libra machine learning toolkit includes implementations of a variety of algorithms for learning and inference with Bayesian networks, Markov networks, and arithmetic circuits. Libra’s strength is exploiting context-specific independence to allow exact inference in models with high treewidth.
Changes:Version 0.3.0 (8/01/2010):
New data structure and functions for Markov networks with factors that are trees, tables, conjunctive features, or sets of conjunctive features.
Added MN support to ACVE, BP, MF, Gibbs, and more.
AC, BN, and MN scoring is now handled by a single program, mscore.
Added mscore utility to convert between .xmod and .bif formats, or to go from .xmod/.bif to .mn (Markov network format).
Added -noac option to aclearnstruct, so that it can be used to learn a Bayesian network that isn’t represented as a circuit.
Added dependency network learner (dnlearn)
Extended tutorial, revised manual, and added more tests.
About: FABIA is a biclustering algorithm that clusters rows and columns of a matrix simultaneously. Consequently, members of a row cluster are similar to each other on a subset of columns and, analogously, members of a column cluster are similar to each other on a subset of rows. A bicluster in a transcriptomic data set is a pair of a gene set and a sample set for which the genes are similar to each other on the samples and vice versa. If multiple pathways are active in a sample, it belongs to different biclusters. If a gene participates in different pathways for different conditions, it belongs to different biclusters, too. Thus, biclusters can overlap. In drug design, for example, researchers want to reveal how compounds affect gene expression; the effects of compounds, however, may be similar only on a subgroup of genes.
About: Implementation of LSTM for biological sequence analysis (classification, regression, motif discovery, remote homology detection). Additionally a LSTM as logistic regression with spectrum kernel is included.
About: Accord.NET is an extension to AForge.NET, a popular .NET framework for computer vision and machine learning. Currently, Accord.NET provides many statistical analysis and processing functions, as well as image processing and computer vision algorithms.
Adding Non-Negative Matrix Factorization, Continuous density Hidden Markov Models and Gaussian Mixture Models;
About: Apache Mahout is an Apache Software Foundation project with the goal of creating both a community of users and a scalable, Java-based framework consisting of many machine learning algorithm […]
Changes:We are pleased to announce release 0.4 of Mahout. Virtually every corner of the project has changed, and significantly, since 0.3. Developers are invited to use and depend on version 0.4 even as yet more change is to be expected before the next release. Highlights include:
* Model refactoring and CLI changes to improve integration and consistency
* New ClusterEvaluator and CDbwClusterEvaluator offer new ways to evaluate clustering effectiveness
* New Spectral Clustering and MinHash Clustering (still experimental)
* New VectorModelClassifier allows any set of clusters to be used for classification
* Map/Reduce job to compute the pairwise similarities of the rows of a matrix using a customizable similarity measure
* Map/Reduce job to compute the item-item-similarities for item-based collaborative filtering
* RecommenderJob has been evolved to a fully distributed item-based recommender
* Distributed Lanczos SVD implementation
* More support for distributed operations on very large matrices
* Easier access to Mahout operations via the command line
* New HMM based sequence classification from GSoC (currently as sequential version only and still experimental)
* Sequential logistic regression training framework
* New SGD classifier
* Experimental new type of NB classifier, and feature reduction options for existing one
* New vector encoding framework for high speed vectorization without a pre-built dictionary
* Additional elements of supervised model evaluation framework
* Promoted several pieces of old Colt framework to tested status (QR decomposition, in particular)
* Can now save random forests and use it to classify new data
* Many, many small fixes, improvements, refactorings and cleanup
About: Pyriel is a Python system for learning classification rules from data. Unlike other rule learning systems, it is designed to learn rule lists that maximize the area under the ROC curve (AUC) instead of accuracy. Pyriel is mostly an experimental research tool, but it’s robust and fast enough to be used for lightweight industrial data mining.
Changes:1.5 Changed CF (confidence factor) to do LaPlace smoothing of estimates. New flag “–score-for-class C” causes scores to be computed relative to a given (positive) class. For two-class problems. Fixed bug in example sampling code (–sample n) Fixed bug keeping old-style example formats (terminated by dot) from working. More code restructuring.