VOCABULARY AND/OR LANGUAGE MODEL TRAINING

摘要

A system (300) comprising means (310) for creating a vocabulary and/or statistical language model (330) from a textual training corpus. The vocabulary and/or language model are used in a pattern recognition system (350), such as a speech recognition system or a handwriting recognition system, for recognising a time-sequential input pattern (352). The system (300) comprises means (335) for determining at least one context identifier and means (332) for deriving at least one search criterion, such as a keyword, from the context identifier. The system further comprises means (334) for selecting documents from a set of documents based on the search criterion. Advantageously, an Internet search engine is used for selecting the documents. Means (336) are used for composing the training corpus from the selected documents.