Skytree Reaches 10 Billion Row Benchmark

Print Friendly, PDF & Email

SkytreeLogo_featureHADOOP SUMMIT 2015 NEWS

Skytree, a leader in enterprise machine learning on big data, announced that it has built a machine learning model with 10 billion rows and 100 columns and ran and completed the first of its kind 1 trillion element benchmark using a sophisticated Gradient Boosted Trees (GBT) model. To celebrate this benchmark, Skytree will kick off a Big Data Challenge at the Hadoop Summit 2015 in San Jose, California. Starting on June 9, conference attendees will be the first to challenge Skytree with their big data problems for a chance to receive two free weeks of Skytree software and data science consultation to assess, analyze and build a model on their business critical use case.

In the era of big data analytics, 10 billion row and trillion element data sets will become increasingly common. We have architected Skytree Infinity to perform no-compromise, no-sub-sampling machine learning-based predictions on massive data sets. To celebrate developing high-accuracy models on trillion element data sets, we are seeking the most compelling massive data sets that we can find to demonstrate the power, scalability and predictive accuracy of Skytree Infinity,” said Ray Villeneuve, CEO of Skytree.

Skytree provides machine learning that automates the most tedious aspects of the data science process. Customers can produce the most accurate models possible in less time and at enterprise-level scale. Skytree’s customers are currently building sophisticated machine-learning models, which are used in production on upwards of half a billion rows. Skytree sought to push past this and designed, developed and executed a benchmark of the Skytree machine learning platform on 10 billion rows of data using 100 features.

Skytree also achieved near-linear scalability, an indication that Skytree can achieve even greater scalability and can analyze even larger data sets without the need to sample the data, which reduces model accuracy. This benchmark demonstrates Skytree’s ability to model the largest and most complex data sets quickly and with increased results.


Sign up for the free insideBIGDATA newsletter.

Speak Your Mind