The LMOffset
parameter determines the offset used for a word hypothesis in a language model. Increasing the offset value tends to reduce the number of word deletion errors at the cost of increasing the number of word insertion errors.
Generally you do not need to change the value of this parameter.
Type: | Long |
Default: | –0.3 |
Recommended Range: | Minimum: -0.5 Maximum: +0.5 |
Required: | No |
Configuration Section: | MyResource (language ID packs only) |
Example: | LMOffset=0.0
|
See Also: |
|