forward-backward algorithm
The "forward-backward algorithm" is a dynamic programming method essential for hidden Markov models, widely used in fields such as speech recognition and bioinformatics.
🇺🇸 US Voice:
🇬🇧 UK Voice:
Definition
C2Hidden Markov Models
(technical, academic)A method to compute the probabilities of hidden states given observed data by making two passes through the sequence, one forward and one backward.
Example
- The forward-backward algorithm allows us to infer the most likely sequence of states in a hidden Markov model.
C2Machine Learning
(technical)An algorithm that calculates the likelihood of observed events by considering all possible hidden state sequences.
Example
- In training a speech recognition system, the forward-backward algorithm is used to estimate model parameters.
Similar
Terms that have similar or relatively close meanings to "forward-backward algorithm":