New features include: Balanced datasets, personalized PageRank, mini-batch optimizer for multilayer perceptron neural networks (and associated pre-processor function), PostgreSQL 10.2 support.
K-nearest neighbors - Added weighted averaging/voting by distance.
Summary - Added more statistics including number of positive, negative, zero values and 95% confidence intervals.
Multilayer perceptron - Added support for one-hot encoded categorical dependent variable for classification.
New feature: Hyperlink-Induced Topic Search (HITS) link analysis algorithm.
k-nearest neighbors (kNN) - Added additional distance metrics, added list of neighbors in output table.
Multlayer perceptron (MLP) - now supports grouping.
Cross validation - Improved the stats reporting in the output table.
Correlation: Improved quality of results by only ignoring a NULL value and not the whole row containing the NULL.
New features include: All Pairs Shortest Path, Weakly Connected Components, Breadth First Search, Mulitple Graph Measures, Stratified Sampling, Train-test split, Multilayer Perceptron and various updates for Apache Top Level Project.
Decision tree and random forest - Allow expressions in feature list, Allow array input for features, Filter NULL dependent values in OOB, Add option to treat NULL as category.
Summary - Allow user to determine the number of columns per run, Improve efficiency of computation time by ~35%.
Sketch - Promote cardinality estimators to top level module from early stage.
MADlib entered incubation in the fall of 2015 and made five releases as an incubating project. Along the way, the MADlib community has worked hard to ensure that the project is being developed according to the principles of the The Apache Way. We will continue to do so in the future as a TLP, to the best of our ability.
Thank you to all who have contributed to the project so far, and we look forward more innovation in machine learning in the future as a TLP!
Downloads for Apache MADlib releases. This also includes links to pre-Apache MADlib releases.