Weka has a package system to extend its functionality, with both official and unofficial packages available, which increases the number of implemented DM methods. It offers four options for DM: command-line interface (CLI), Explorer, Experimenter, and Knowledge flow. While Weka isn’t aimed specifically at Hadoop users and Big Data processing, it can be used with Hadoop thanks to a set of wrappers produced for the most recent versions of Weka3. At the moment, it still does not support Apache Spark, but only MapReduce. Clojure [Clojure] users can also leverage Weka, thanks to the Clj-ml library [Clj-ml]. Related to Weka, Massive Online Analysis (MOA) is also a popular open source framework written in Java for data stream mining, while scaling to more demanding larger-scale problems.

Strong points

• General purpose, well-maintained, involving wide set of algorithms with learning schemes, models and algorithms.
• It comes with GUI and API-oriented.
• Supports standard DM tasks, including feature selection, clustering, classification, regression and visualization.
• Very popular ML tool in the academic community.

Weak points

• Limited for Big Data, text mining, and semi-supervised learning.
• Weak for sequence modelling e.g. time-series.

Return to Contemt