

The Apache Software Foundation said today that it’s designating open-source machine learning software first developed by Salesforce.com Inc. as its latest top-level project.
The project is called Apache PredictionIO, and is described as an open-source machine learning server that lets developers and data scientists create predictive engines and services to perform machine learning tasks. The idea is to democratize machine learning by providing a full software “stack” for creating intelligent applications “without having to cobble together the underlying technologies,” PredictionIO founder Simon Chan, now the senior director of Salesforce’s Einstein artificial intelligence initiative, wrote in a blog post.
Salesforce quietly acquired PredictionIO back in February 2016, and first used the startup’s technology to create an “agenda builder” application that suggested relevant sessions to those attending technical conferences. That app worked by gauging user’s preferences along with those of other attendees at the shows in order to build personalized agendas.
Salesforce has built a few other applications using PredictionIO since then, including apps for predicting drop out rates at colleges and customized predictive models for banks.
PredictionIO helps make machine learning-based apps easier to build thanks to its “template gallery,” which offers tools such as classification, clustering, natural language, recommendation and regression engines. Developers can help themselves to these engines by downloading them in order to save a significant amount of time. With PredictionIO, recommendation engines, which normally take months to build, can be created within just a “couple of weeks,” Chan said.
Salesforce’s hope is that PredictionIO will help to speed up the adoption of machine learning, which it says is central to data analytics and artificial intelligence.
PredictionIO is bundled with tools including the Apache Spark big data processing framework, MLLib, Hbase and Elasticsearch, among other technologies. It can be downloaded from GitHub.
THANK YOU