Training own model in opennlp

问题

I am finding it difficult to create my own model openNLP. Can any one tell me, how to own model. How the training shouls be done.

What should be the input and where the output model file will get stored.

回答1:

https://opennlp.apache.org/docs/1.5.3/manual/opennlp.html

This website is very useful, shows both in code, and using the OpenNLP application to train models for all different types, like entity extraction and part of speech etc.

I could give you some code examples in here, but the page is very clear to use.

Theory-wise:

Essentially you create a file which lists the stuff you want to train

eg.

Sport [whitespace] this is a page about football, rugby and stuff

Politics [whitespace] this is a page about tony blair being prime minister.

The format is described on the page above (each model expects a different format). once you have created this file, you run it through either the API or the opennlp application (via command line), and it generates a .bin file. Once you have this .bin file, you can load it into a model, and start using it (as per the api in the above website).

回答2:

First you need to train the data with the required Entity.

Sentences should be separated with new line character (\n). Values should be separated from and tags with a space character.
Let's say you want to create medicine entity model, so data should be something like this:

<START:medicine> Augmentin-Duo <END> is a penicillin antibiotic that contains two medicines - <START:medicine> amoxicillin trihydrate <END> and 
<START:medicine> potassium clavulanate <END>. They work together to kill certain types of bacteria and are used to treat certain types of bacterial infections.

You can refer a sample dataset for example. Training data should have at least 15000 sentences to get the better results.

Further you can use Opennlp TokenNameFinderTrainer. Output file will be in the .bin format.

Here is the example: Writing a custom NameFinder model in OpenNLP

For more details, refer the Opennlp documentation

回答3:

Perhaps this article will help you out. It describes how to do TokenNameFinder training from data extracted from Wikipedia...

nuxeo - blog - Mining Wikipedia with Hadoop and Pig for Natural Language Processing

回答4:

Copy the data in data and run below code to get your own mymodel.bin .

Can refer for data=https://github.com/mccraigmccraig/opennlp/blob/master/src/test/resources/opennlp/tools/namefind/AnnotatedSentencesWithTypes.txt

public class Training {
       static String onlpModelPath = "mymodel.bin";
       // training data set
       static String trainingDataFilePath = "data.txt";

       public static void main(String[] args) throws IOException {
                       Charset charset = Charset.forName("UTF-8");
                       ObjectStream<String> lineStream = new PlainTextByLineStream(
                                                       new FileInputStream(trainingDataFilePath), charset);
                       ObjectStream<NameSample> sampleStream = new NameSampleDataStream(
                                                       lineStream);
                       TokenNameFinderModel model = null;
                       HashMap<String, Object> mp = new HashMap<String, Object>();
                       try {
                              //         model = NameFinderME.train("en","drugs", sampleStream, Collections.<String,Object>emptyMap(),100,4) ;
                                       model=  NameFinderME.train("en", "drugs", sampleStream, Collections. emptyMap());
                       } finally {
                                       sampleStream.close();
                       }
                       BufferedOutputStream modelOut = null;
                       try {
                                       modelOut = new BufferedOutputStream(new FileOutputStream(onlpModelPath));
                                       model.serialize(modelOut);
                       } finally {
                                       if (modelOut != null)
                                                       modelOut.close();
                       }
       }
}

来源：https://stackoverflow.com/questions/11204352/training-own-model-in-opennlp

标签

file

model

opennlp