The DoSmoothing
parameter determines whether to use smoothing for custom language models.
During smoothing, the probability estimates obtained from training text are discounted to avoid overfitting the training data. In some situations you might want to train a model from text that is an exact transcription of the speech to be recognized. In this case you disable smoothing, which might improve recognition.
Type: | Boolean |
Default: | True |
Required: | No |
Configuration Section: | lmbuild module |
Example: | DoSmoothing=True
|
See Also: | DoSmoothing (action parameter) |
|