Two-stage Thai Misspelling Correction based on Pre-trained Language Models.
Idhibhat PankamPeerat LimkonchotiwatEkapol ChuangsuwanichPublished in: JCSSE (2023)
Keyphrases
- language model
- pre trained
- language modeling
- n gram
- training data
- document retrieval
- language modelling
- information retrieval
- training examples
- word segmentation
- probabilistic model
- retrieval model
- test collection
- query expansion
- speech recognition
- control signals
- neural network
- document ranking
- statistical language models
- relevance model
- small number
- smoothing methods
- language models for information retrieval
- supervised learning
- feature extraction