By comparing a single document to a set of two or more other documents, StyleScorer evaluates how closely they match in terms of writing style. Using statistical language modeling, natural language processing (NLP), forensic linguistics and neural networks, StyleScorer gives each document a score which determines how closely it matches the data-driven style guidelines.
Clients can then determine the suitability of the source and translated content for publishing. StyleScorer is language agnostic. The documents being compared must be in the same language; however, there is no restriction on the language of the content. As long as the documents you’re scoring adhere to a set style (or a style guide), then the scores will give a good indication of whether the content is in line with your overall brand voice and meets quality requirements.
Welocalize StyleScorer Features
• Delivered as Part of Welocalize’s weMT Program
• Forms Part of QA + Linguistic Review Workflow for MT + Human Translation
• Documents Scored on an “As Needed” Basis
• Scores Source + Translated Documents Against Training Content
• Can Be Used to Determine Effectiveness of MT Engines
• Determines Quality of MT Output for Post-Editing
• Language Agnostic as Long as Documents in Same Language
• Easy Connection into weMT Program
• Metadata for Business Intelligence
• Suitable for All Types of Content
StyleScorer enhances QA and quality estimation, as part of the overall localization and MT workflow. Use of linguistic tools enables more content to be translated while maintaining brand voice and quality standards and without significantly increasing costs. By using Welocalize StyleScorer as part of your weMT program, content published is more consistent and truly represents a global voice across all products and divisions.
Clients can use StyleScorer as part of their Welocalize weMT program to help streamline linguistic review workflows. StyleScorer output can determine whether more MT engines are required for different types of content. If you run StyleScorer on raw MT output, the scores can be used to rank which documents need the most post-editing to bring them in line with the style guidelines. It can also be used to spot-check human and post-edited translations, providing statistical reporting and analytics. Even if you don’t have a formal, company-wide style-guide, StyleScorer will still work as long as the training documents are identified as belonging to a cohesive group.
Welocalize weMT StyleScorer helps you manage your brand voice and tone across your entire global organization.