Properly reconstruct sequence of states in Viterbi algorithm #4
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
added 7 commits
September 14, 2022 15:13
…bability In a latter commit, we will add additional instructions in each branch to track the sequence of intermediate states.
We will need to disambiguate between a single weather state and a list of weather states in a latter commit.
Before this commit we were just selecting the state with max probability at each time step, but this would not properly rebuild the sequence of states leading to the final most likely state. Note that this changes the result.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Hi, thanks a lot for the tutorial, I found it very intuitive.
However there is a mistake at the end of the part concerning the Viterbi algorithm. It is not possible to simply select the state with highest probability at each step. Doing this will not reconstruct properly the sequence that lead to the most likely final state.
Instead, we have to memorize all the best intermediate states and how they connect to each other. Then when we have selected our final state, we just have to "backtrack" the series of states that lead to it.