This is a tutorial for Wapiki, a tool for CRF implementation.
The proper way to use the materials in the tutorial:
Step 1. Read the text book. Klinger, Roman, and Katrin Tomanek. Classical probabilistic models and conditional random fields. TU, Algorithm Engineering, 2007. Link: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.645.5543&rep=rep1&type=pdf
Alternatively, Sutton, Charles, and Andrew McCallum. "An introduction to conditional random fields." Foundations and Trends® in Machine Learning 4, no. 4 (2012): 267-373. Link of a short version: https://publist.ist.ac.at/attachments/0000/0292/crf-tutorial.pdf
For readers who'd like to know more about parameter optimization. Viterbi algorithm on HMM is a MUST-TO-READ: Link: http://www.robots.ox.ac.uk:5000/~vgg/rg/papers/hmm.pdf
For undertanding CRF in a concret mathmatical way, check the slide 《图模型，ME，HMM和CRF》in https://hzaubionlp.com/nlp4underg/.
Step 2. Follow the experiment tutorial. CRF实验.pdf The tutorial is offered by @KaiyinZhou.
Step 3. Read the pad.pdf for detail about the patten design.
Patterns link: https://wapiti.limsi.fr/manual.html#patterns
Aknowledgement: Thank Pierre Zweigenbaum from LIMSI offered tutorial of WAPITI, a CRF tool, in 2017 to HZAU-BioNLP lab here. Thank all of the audiences in BioNLP-KD courese (since 2016) and NLP course (Since 2020) who join the algorithmic discussions.