In Natural Language Process (NLP), POS-tagger is an essential process, which helps to understand the Natural Language queries for computer. In abstract an implemented model is trained with tags, linguistic corpus and generates trained tagger model.
The Stanford POS-tagger is one of the most popular tagger. It’s been developed, optimized and pruned for more than 10 years. The POS-tagger can be downloaded from this following site: nlp.stanford.edu
Stanford is matured framework where it allows to train the models with our own corpus.
To train a model, navigate to Stanford POS-tagger package, which you downloaded. The standford-postagger.jar is instructed using a PROPS file. First generate the property file which includes the template:
This will generate the default property file with all the details as comments. You can modify the default option or use the default option to train the models.
These are the modification I have done to my default property file:
Training Data set for a specific domain (Eg: Retail Domain)