There are a number of other interesting projects out there which have some features in common with Palladium. In the following, we will mention a selection.
Infrastructure for data analysis (PaaS). A UI for running and examining experiments is provided. Experiments can be run in parallel and notification mechanisms can be set up. Models can be deployed as web services (referring to DOMINO’s documentation, the overhead for the HTTP server is about 150ms) and model updates can be scheduled. Supports Python, R, Julia, Octave, and SAS models. Commercial.
ML server based on Hadoop / Spark. Two engine templates for Apache Spark MLlib are provided for setting up a recommendation engine or a classification engine. It also allows for gathering new events. Supports Spark MLlib and Scala models (no support for Python, R, Julia). Open source.
Tool to support experiments performed with scikit-learn. It allows to run various settings on different test sets and to get a summary of the test’s results. It does not aim at exposing models as web services. Supports Python models (no support for R or Julia). Open source.
Platform for managing predictive models in production environments (PaaS). A command line tool and GUI are available for model management. It also provides a Load Balancer and can automatically scale the servers as needed. Supports Python and R models. Commercial.