named_entity_recognition

Why we included this project

This repository works through four approaches to Chinese named entity recognition on the same dataset, which makes the trade-offs easy to see. Each model, from the plain HMM up to the BiLSTM+CRF stack, has its own file with training and decoding code, and the README explains the theory behind each one and what the next model fixes, so the progression from probabilistic to neural taggers reads clearly. It ships with a resume corpus labeled in BIOES format, and running main.py trains all four and prints precision, recall, F1, and a confusion matrix, so you can compare them side by side on identical data. If you are learning sequence labeling or want a compact reference before building your own Chinese NER pipeline, this is a useful starting point.