What is NGramJ?
Getting Started Download Instructions Decide Which Type Use CNgram Use NGramJ
How Does it Work?
Contact
How to Contribute?
Developer Information
Other Information
|
Usage is very simple.
- Unpack
cngram.jar from the binary download.
- Type
cngram.jar on the commandline. This is an self executing jar and this works on JRE installations from version 1.2. If your JRE is misconfigured use the equivalent call
java -jar cngram.jar
- This command has several options, which are essentially self explaining.
- Note A: Some options are rather for testing than for usage.
- Most common call is
cngram.jar -lang2 YOUR_FILENAME FILE_ENCODING which calls the new enhanced language determination. (-lang uses the old algorithm, -lang2b calls the new algorithm twice to demonstrate speedup.)
- Note B: File encoding for input files with CNgram is optional, if you don't give on explicitely, ISO-8859-1 is used.
- When a metrics class is requested, you have to provide the full qualified classname of a class implementing
NGramMetric , thus describing a way how different ngram profiles are compared. Note, the RawMetric seems to be the most common used.
|