Bayes Server versions 6 and later support distributed processing on platforms such as Apache Spark and Apache Hadoop, including support for distributed learning of time series models.

A few examples of models you can build using distributed Bayes Server.

• Distributed Bayesian networks
• Distributed Mixture models (clustering)
• Distributed Time Series models
• Distributed Naive Bayes
• Distributed Mixture of Naive Bayes
• Distributed Hidden markov models
• Distributed Dynamic Bayesian networks
• Distributed Mixtures of Time Series
• Distributed anomaly detection models
• and many more...

The distributed implementation in Bayes Server is agnostic to the distributed platform, and works on platforms other than Hadoop and Spark.

In addition both the Java and .NET APIs support distributed processing, and can also be called by derivatives such as Scala.

Bayes Server 6 only supports distributed parameter learning, so the structure of the model must be known in advance. In 9 out of 10 cases this is not a problem. Expect to see distributed structural learning in a future release.

Apache Spark

These code samples demonstrate how to build Bayes Server models such as probabilistic cluster models and time series models on Apache Spark.

Big data consultancy

We also provide big data consultancy. To find out more please get in touch, and a consultant will contact you.