Three Methods for eDiscovery Document Prioritization: Comparing and Contrasting Keyword Search with Concept Based and Support Vector Based “Technology Assisted Review-Predictive Coding” Platforms


This paper explores the three primary approaches eDiscovery professionals use to systematically prioritize documents that are being considered for review:

1) Keyword Search

2) TAR-PC based on Categorization with Conceptual Search Technology

3) TAR-PC based on an Active Machine Learning approach using a Support Vector Machine

Each of these three approaches has its own unique set of advantages and disadvantages and, when used in conjunction with careful documentation and appropriate iteration, can survive a challenge by a party’s opponent in litigation. The optimal workflow will likely require a hybrid approach applying two or perhaps all three of these approaches to reach the best results.

The implementation of Technology Assisted Review – Predictive Coding found in concept based Relativity Assisted Review and Support Vector based Equivio Relevance yields greater results than just using keywords search. These systems can return more potentially content relevant documents without the limitations of Boolean logic. However, a well-constructed keyword search can be more effective for certain conditions such as dates or proper names. Putting these technologies together with proper workflow, methodology and documentation will deliver the most effective results and enhance their defensibility.

Perhaps more important than the technology itself is how the skilled eDiscovery professional -- who understands these systems, their limitations and strengths -- combines them into the optimal workflow best suited to meet the overall requirements for the given project. In the end, the defensibility of these approaches is based on the workflow (the process) and not the technology itself.

LOADING PDF: If there are any problems, click here to download the file.

DISCLAIMER: Because of the generality of this update, the information provided herein may not be applicable in all situations and should not be acted upon without specific legal advice based on particular situations.

© Tom Groom, D4 LLC | Attorney Advertising

Written by:


D4 LLC on:

Readers' Choice 2017
Reporters on Deadline

"My best business intelligence, in one easy email…"

Your first step to building a free, personalized, morning email brief covering pertinent authors and topics on JD Supra:

Sign up to create your digest using LinkedIn*

*By using the service, you signify your acceptance of JD Supra's Privacy Policy.

Already signed up? Log in here

*With LinkedIn, you don't need to create a separate login to manage your free JD Supra account, and we can make suggestions based on your needs and interests. We will not post anything on LinkedIn in your name. Or, sign up using your email address.