Using Pre-Coded Documents to Train Relativity Assisted Review

by Constantine Pappas on February 28, 2014

Analytics & Assisted Review , Product Spotlight

Introduced in Relativity 8, pre-coded rounds take documents that have already been coded and bring them into a Relativity Assisted Review project. When case teams take advantage of this option, documents with designations already provided can be used to train the system. Training the system in this way reduces the need to repeat work, and therefore saves review time.

It’s not uncommon to begin a project in a rush and then realize you need to take a different approach. Assisted Review allows you to leverage analytics to code fewer documents, so it’s especially helpful in very large cases. When you’ve already begun manual review but decide Assisted Review might be a more effective option, pre-coded rounds are the best option to bring those coded values into the project right from the start. That said, it’s key to remember that those documents need to be reviewed based on the content of the individual document and not the family. Documents coded based on the content of their families can train the system incorrectly.

Pre-coded sets are a great way to use old control sets, too. When the dynamic of the data set changes—often because new documents have been added—a new control set should be created. To prevent wasted work, the old control set documents can then be used to train the system with a pre-coded round.

To learn more about pre-coded rounds in Assisted Review, check out our documentation and, as always, feel free to contact us if you have any questions.

Posted by Constantine Pappas.


Post a Comment

Required Field