I am new to Kalman filter and parsers. I understand that it's suitable for prediction/estimation on the go, when you have something to measure (not only the model), and when changes always occur in the measurement or model.

However, for the case of finite automata, pushdown automata, or grammars, usually the automata or grammars do not change. However, I found papers combining Kalman filter and fuzzy automata, or Kalman filter and cellular automata.

Does this means Kalman filter is not suitable to be used in finite/pushdown automata or context-free grammars, unless those models are equipped with fuzziness, probability, or weights?

Is there any parser or parsing method using Kalman filter? I would appreciate it very much if you share your wisdom about why it's used or not used.

  • 2
    $\begingroup$ Traditional parsing is deterministic and the algorithms have perfect knowledge of the input. I'm not sure how or why you'd want to insert Kalman filters into the process. $\endgroup$ – adrianN Jan 30 '17 at 8:20
  • $\begingroup$ @adrianN, how about a non-deterministic finite/pushdown automata, where you can go to different states while reading the same input? $\endgroup$ – kate Jan 30 '17 at 8:25
  • $\begingroup$ Also in nondeterministic automata the result ("could parse, or could not parse") is determined only by the input. Nondeterminism is different from randomization. $\endgroup$ – adrianN Jan 30 '17 at 8:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.