• Engineer meticulously
• Transform an industry
Predictive coding is a machine-learning technology that can be trained, by way of examples, to imitate the decisions of a subject-matter expert. Predictive coding is used in e-discovery to focus on key documents, to cull raw collections, and to prioritize review. In information governance, predictive coding is used to classify records and identify junk. To be effective, it’s critical that predictive coding applications be based on supervised classification technologies.
Equivio’s Themes application is a new-generation clustering technology. Themes allows the intuitive exploration of data populations. Themes is used when someone needs to make sense of a collection of documents. For example, in e-discovery, for case planning and strategizing (what early case assessment should be about). Or in investigations, because of its ability to guide you through a collection where you’re not even sure what you’re looking for.
Email threading reconstructs email conversations from unstructured data populations. Threading enables the review of emails within the context of the thread from which the email originated. It’s used by investigations, e-discovery review and regulatory processes when people need to understand who was speaking to whom about what. And it’s used to slash document review costs. For legal and compliance applications, the defensibility of email threading is dependent on the analysis of email content rather than the interpretation of metadata.
Near-duplicate analysis refers to the grouping of very similar documents. In contrast to clustering, which groups documents that are related by subject, near-duping algorithms group documents whose textual content is almost but not quite identical. Near-duplicate analysis reduces the effective volume of document populations, simplifying complexity and eliminating noise from large data sets. The technology is used to accelerate investigations and document review processes.