The key to delivering next generation translation is the creation of multi-lingual dictionaries of your organization's specific terminology.
Initially a source language dictionary is created. This is usually then sent to a terminologist for translation.
The dictionary created by Translution consists of two types of data:
- Terms (including compound terms) which are not in the standard Translution dictionaries or which have specific meaning in your industry
- Terms that you do not want translated such as names of people, places, companies and product names
The dictionary is created from your own data using statistical and specialist linguistic tools.
This data can be anything that is relevant to your business, for instance, operations manuals, quality manuals, product information, contracts, contacts, standard proposals, correspondence, your website etc.
Our experience shows that a statistical approach is the most comprehensive and cost effective way to rapidly build your organization's specific terminology. Typical sizes of dictionaries created are between 500 and 3000 terms and as many Do Not Translate terms as you want.
Once built, your dictionaries provide you with a key business resource that has many uses in addition to improving machine translation quality. Examples are in improving consistency of authoring tasks, improving the accuracy of human translation and helping with the induction of new employees.
The dictionary creation process typically takes between 4 and 6 weeks and consists of 3 distinct stages:
Stage 1 Collect the source data
Translution creates a corpus of mono-lingual data derived from your existing electronic documents and databases.
Stage 2 Identify source terms
Translution extract the base form (for example - the singular form) of terms from the corpus using a range of specialist processes and software tools. This produces a spreadsheet of terms such as nouns, proper nouns, compound terms and acronyms suitable for reviewing.
Stage 3 Dictionary creation
Once reviewed and approved by you, source terms are translated into the required languages by specialist terminologists to create a multi lingual dictionary. The final dictionary is then published.