CRF Training Algorithm

From GM-RKB
Jump to navigation Jump to search

A CRF Training Algorithm is a graphical model training algorithm that accepts a CRF model family and can be applied by a CRF training system (to solve a CRF training task to produce a CRF structure).



References

2016

  • http://www.datasciencecentral.com/profiles/blogs/conditional-random-fields-crf-short-survey
    • QUOTE: The time complexity of the training process is large enough: [math]\displaystyle{ \begin{equation*} O(mNTQ2nS), \end{equation*} }[/math], where:
      • m is the number of training iterations
      • N is the number of training data sequences
      • T is the average length of training sequences
      • Q is the number of class labels
      • n is the number of CRF features
      • S is the searching time of the optimization algorithm (for example, L-BFGS algorithm, which is considered good for this).
    • In practical implementation, the computational time is often larger due to many other operations like numerical scaling, smoothing etc.

2010

2006

2003