This documentation introduces Verdict’s core concepts and operations. Verdict is developed to provide fast and resource-efficient analytics, working on top of existing engines (thus, no modifications to the engines).
Unlike other systems, Verdict predicts the answers (instead of simply aggregating individual values) by exploiting statistical properties of analytical workloads. These predictions are performed almost instantly and come with tight accuracy guarantees. In many cases, you get more than 99% accurate answers in less than a second.
Everything in this project is under the Apache License.