Massive Online Analysis Explained

Massive Online Analysis (MOA) is a free open-source software project specific for data stream mining with concept drift. It is written in Java and developed at the University of Waikato, New Zealand.^[1]

Description

MOA is an open-source framework software that allows to build and run experimentsof machine learning or data mining on evolving data streams. It includes a set of learners and stream generators that can be used from the Graphical User Interface (GUI), the command-line, and the Java API.MOA contains several collections of machine learning algorithms:

• Classification
  • Bayesian classifiers
    • Naive Bayes
    • Naive Bayes Multinomial
  • Decision trees classifiers
    • Decision Stump
    • Hoeffding Tree
    • Hoeffding Option Tree
    • Hoeffding Adaptive Tree
  • Meta classifiers
    • Bagging
    • Boosting
    • Bagging using ADWIN
    • Bagging using Adaptive-Size Hoeffding Trees.
    • Perceptron Stacking of Restricted Hoeffding Trees
    • Leveraging Bagging
    • Online Accuracy Updated Ensemble
  • Function classifiers
    • Perceptron
    • Stochastic gradient descent (SGD)
    • Pegasos
  • Drift classifiers
    • Self-Adjusting Memory^[2]
    • Probabilistic Adaptive Windowing
  • Multi-label classifiers^[3]
  • Active learning classifiers ^[4]
• Regression
  • FIMTDD^[5]
  • AMRules^[6]
• Clustering^[7]
  • StreamKM++
  • CluStream
  • ClusTree
  • D-Stream
  • CobWeb.
• Outlier detection^[8]
  • STORM
  • Abstract-C
  • COD
  • MCOD
  • AnyOut^[9]
• Recommender systems
  • BRISMFPredictor
• Frequent pattern mining
  • Itemsets^[10]
  • Graphs^[11]
• Change detection algorithms^[12]

These algorithms are designed for large scale machine learning, dealing with concept drift, and big data streams in real time.

MOA supports bi-directional interaction with Weka (machine learning). MOA is free software released under the GNU GPL.

See also

• ADAMS Workflow: Workflow engine for MOA and Weka (machine learning)
• Streams: Flexible module environment for the design and execution of data stream experiments
• Weka (machine learning)
• Vowpal Wabbit
• List of numerical analysis software

External links

• MOA Project home page at University of Waikato in New Zealand
• SAMOA Project home page at Yahoo Labs

Notes and References

1. Bifet . Albert . Holmes . Geoff . Kirkby . Richard . Pfahringer . Bernhard . MOA: Massive online analysis . The Journal of Machine Learning Research . 99 . 1601–1604. 2010.
2. Losing. Viktor. Hammer. Barbara. Wersing. Heiko. Tackling heterogeneous concept drift with the Self-Adjusting Memory (SAM). Knowledge and Information Systems. 54. 171–201. 2017. 0885-6125. 10.1007/s10115-017-1137-y. 29600755.
3. Read. Jesse. Bifet. Albert. Holmes. Geoff. Pfahringer. Bernhard. Scalable and efficient multi-label classification for evolving data streams. Machine Learning. 88. 1–2. 2012. 243–272. 0885-6125. 10.1007/s10994-012-5279-6. 14676146. free.
4. Zliobaite. Indre. Bifet. Albert. Pfahringer. Bernhard. Holmes. Geoffrey. Active Learning With Drifting Streaming Data. IEEE Transactions on Neural Networks and Learning Systems. 25. 1. 2014. 27–39. 2162-237X. 10.1109/TNNLS.2012.2236570. 24806642. 14687075.
5. Ikonomovska. Elena. Gama. João. Džeroski. Sašo. Learning model trees from evolving data streams. Data Mining and Knowledge Discovery. 23. 1. 2010. 128–168. 1384-5810. 10.1007/s10618-010-0201-y. 7114108.
6. Book: Almeida. Ezilda. Advanced Information Systems Engineering. Ferreira. Carlos. Gama. João. Adaptive Model Rules from Data Streams. 8188. 2013. 480–492. 0302-9743. 10.1007/978-3-642-40988-2_31. Lecture Notes in Computer Science. 978-3-642-38708-1. 10.1.1.638.5472.
7. Book: Kranen. Philipp. 2010 IEEE International Conference on Data Mining Workshops. Kremer. Hardy. Jansen. Timm. Seidl. Thomas. Bifet. Albert. Holmes. Geoff. Pfahringer. Bernhard. Clustering Performance on Evolving Data Streams: Assessing Algorithms and Evaluation Measures within MOA. 2010. 1400–1403. 10.1109/ICDMW.2010.17. 978-1-4244-9244-2. 2064336.
8. Book: Georgiadis. Dimitrios. Proceedings of the 2013 international conference on Management of data - SIGMOD '13. Kontaki. Maria. Gounaris. Anastasios. Papadopoulos. Apostolos N.. Tsichlas. Kostas. Manolopoulos. Yannis. Continuous outlier detection in data streams. 2013. 1061. 10.1145/2463676.2463691. 9781450320375. 1886134.
9. Book: Assent. Ira. Database Systems for Advanced Applications. Kranen. Philipp. Baldauf. Corinna. Seidl. Thomas. AnyOut: Anytime Outlier Detection on Streaming Data. 7238. 2012. 228–242. 0302-9743. 10.1007/978-3-642-29038-1_18. Lecture Notes in Computer Science. 978-3-642-29037-4.
10. Quadrana. Massimo. Bifet. Albert. Gavaldà. Ricard. An Efficient Closed Frequent Itemset Miner for the MOA Stream Mining System. Frontiers in Artificial Intelligence and Applications. 256. Artificial Intelligence Research and Development. 2013. 203. 10.3233/978-1-61499-320-9-203.
11. Book: Bifet. Albert. Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining - KDD '11. Holmes. Geoff. Pfahringer. Bernhard. Gavaldà. Ricard. Mining frequent closed graphs on evolving data streams. 2011. 591. 10.1145/2020408.2020501. 9781450308137. 10.1.1.297.1721. 8588858.
12. Book: Bifet. Albert. Advances in Intelligent Data Analysis XII. Read. Jesse. Pfahringer. Bernhard. Holmes. Geoff. Žliobaitė. Indrė. CD-MOA: Change Detection Framework for Massive Online Analysis. 8207. 2013. 92–103. 0302-9743. 10.1007/978-3-642-41398-8_9. Lecture Notes in Computer Science. 978-3-642-41397-1.