What is this?
Apache PredictionIO is an open source machine learning server that lets developers and data scientists collect events, deploy algorithms, evaluate them, and query predictive results through REST APIs.
Apache PredictionIO is an open source machine learning server that lets developers and data scientists collect events, deploy algorithms, evaluate them, and query predictive results through REST APIs.Is this a library I import, or a standalone server I run?It is a standalone server built on top of Hadoop, HBase, Elasticsearch and Spark, implementing a Lambda Architecture, which your app talks to over REST APIs rather than embedding as code.
It is a standalone server built on top of Hadoop, HBase, Elasticsearch and Spark, implementing a Lambda Architecture, which your app talks to over REST APIs rather than embedding as code.How do I get it running?You can install Apache PredictionIO from a binary or source release, or run it with Docker, then use one of the provided engine templates (Recommendation, Similar Product, Classification) to get a Quick Start going.
You can install Apache PredictionIO from a binary or source release, or run it with Docker, then use one of the provided engine templates (Recommendation, Similar Product, Classification) to get a Quick Start going.What infrastructure does it need to run?It relies on a stack of other big-data services: Hadoop, HBase (or another supported DB), Elasticsearch and Spark, rather than running as a single lightweight process.
It relies on a stack of other big-data services: Hadoop, HBase (or another supported DB), Elasticsearch and Spark, rather than running as a single lightweight process.Is this project still maintained?No, the repository is archived, so it no longer receives active development or updates from the Apache project.
No, the repository is archived, so it no longer receives active development or updates from the Apache project.Who is this built for?It targets developers, data scientists, and end users who want to add predictive features like recommendations or classification to their apps without building the ML infrastructure from scratch.
It targets developers, data scientists, and end users who want to add predictive features like recommendations or classification to their apps without building the ML infrastructure from scratch.What kind of engine templates come with it?The README highlights Quick Start guides for a Recommendation Engine Template, a Similar Product Engine Template, and a Classification Engine Template.
The README highlights Quick Start guides for a Recommendation Engine Template, a Similar Product Engine Template, and a Classification Engine Template.What architecture pattern does it implement?It implements a Lambda Architecture, combining batch and real-time processing across its underlying big-data services.
It implements a Lambda Architecture, combining batch and real-time processing across its underlying big-data services.Where is the full documentation?Docs are included in the repo under docs/manual and also hosted publicly at predictionio.apache.org.
Docs are included in the repo under docs/manual and also hosted publicly at predictionio.apache.org.Should I start a new project with this today?Since the repository is archived, it's best suited for studying its architecture or maintaining existing deployments rather than starting new production systems.
Since the repository is archived, it's best suited for studying its architecture or maintaining existing deployments rather than starting new production systems.What are similar projects?spark, trino, elasticsearch, OpenSearch, presto, dask, cube, logstash
spark, trino, elasticsearch, OpenSearch, presto, dask, cube, logstashWhich topics is it in?recommendation-system, hadoop, scala, elasticsearch, big-data, analytics, data-science, distributed-systems, deprecated, rest-api, database, framework
recommendation-system, hadoop, scala, elasticsearch, big-data, analytics, data-science, distributed-systems, deprecated, rest-api, database, framework