python 3 nltk numpy

• run main.py to get predicted results of HMM or MEMM
• run run_hmmem.py to get predicted results of HMM(em)
• run crf++ with command lines, with the train/test files in data/; predicted results can be saved in data/pred_results/
• user scripts/conlleval_rev.pl to evaluate files in data/pred_results/

• put new data in data/orig/ with format:

不/d 忘/v 藏北/s 人民/n 的/u 拉萨/ns 市民/n （/w 图片/n ）/w

• run data/transform2conll.py to transform data/orig files to seg/ner/pos files with crf++ format

• blog link (to be added)
• https://github.com/tostq/Easy_HMM
• https://spaces.ac.cn/archives/3922
• https://github.com/yh1008/MEMM
• https://cocoxu.github.io/courses/5525_slides_spring17/15_more_memm.pdf
• https://taku910.github.io/crfpp/