POS Tagger with Stanza pipeline

[dxv2k]

DATA ANNOTATION: Penn Treebank 2 format

WORK FLOW with Stanza:
Document (in CoNLLU format or perform conversion ) -> Setence Segmentation -> Tokenize and Multi-word Tokenize-> POS Tagging

NOTICE:

• We work with English so no need for multi-word tokenization (MWT)
• In order to use Stanza, data format must be in CoNLLU
• Use pre-trained tokenization and POS Tagger to compare with Viterbi Algorithm (XPOS field)
• Neural pipeline in Stanza: Maximum entropy cyclic dependency network

Essential libraries and other components will be listed later.