# Viterbi algorithm

**Viterbi algorithm** is a algorithm for finding the most sequence of hidden states—called the **Viterbi path**—that results in a sequence of observed events, especially in the context of and .

The algorithm has found universal application in decoding the convolutional codes used in both and digital cellular, modems, satellite, deep-space communications, and wireless LANs. It is now also commonly used in , , , , , and . For example, in (speech recognition), the acoustic signal is treated as the observed sequence of events, and a string of text is considered to be the “hidden cause” of the acoustic signal. The Viterbi algorithm finds the most likely string of text given the acoustic signal.

## Contents

## History

The Viterbi algorithm is named after , who proposed it in 1967 as a decoding algorithm for over noisy digital communication links. It has, however, a history of , with at least seven independent discoveries, including those by Viterbi, , and .

“Viterbi path” and “Viterbi algorithm” have become standard terms for the application of dynamic programming algorithms to maximization problems involving probabilities.

## Extensions

A generalization of the Viterbi algorithm, termed the *max-sum algorithm* (or *max-product algorithm*) can be used to find the most likely assignment of all or some subset of in a large number of , e.g. , and . The latent variables need in general to be connected in a way somewhat similar to an HMM, with a limited number of connections between variables and some type of linear structure among the variables. The general algorithm involves *message passing* and is substantially similar to the algorithm (which is the generalization of the ).

With the algorithm called iterative Viterbi decoding one can find the subsequence of an observation that matches best (on average) to a given . This algorithm is proposed by Qi Wang et al. to deal with turbo code. Iterative Viterbi decoding works by iteratively invoking a modified Viterbi algorithm, reestimating the score for a filler until convergence.

An alternative algorithm, the , has been proposed recently. For many codes of practical interest, under reasonable noise conditions, the lazy decoder (using Lazy Viterbi algorithm) is much faster than the original Viterbi decoder (using Viterbi algorithm). This algorithm works by not expanding any nodes until it really needs to, and usually manages to get away with doing a lot less work (in software) than the ordinary Viterbi algorithm for the same result—however, it is not so easy to parallelize in hardware.

## Pseudocode

This algorithm generates a path , which is a sequence of states that generate the observations ( being the count of observations (observation space, see below)).

Two 2-dimensional tables of size are constructed:

- Each element of stores the probability of the most likely path so far with that generates .
- Each element of stores of the most likely path so far for

The table entries are filled by increasing order of .

- , and
- ,

with and as defined below. Note that does not need to appear in the latter expression, as it’s non-negative and independent of and thus does not affect the argmax.

- INPUT

- The
- the
- an array of initial probabilities such that stores the probability that
- a sequence of observations such that if the observation at time is
- of size such that stores the of transiting from state to state
- of size such that stores the probability of observing from state

- OUTPUT

- The most likely hidden state sequence

functionVITERBIforeach statedoend forforeach observationdoforeach statedoend forend forfordoend forreturnend function

- EXPLANATION

Suppose we are given a (HMM) with state space , initial probabilities of being in state and transition probabilities of transitioning from state to state . Say we observe outputs . The most likely state sequence that produces the observations is given by the recurrence relations:

Here is the probability of the most probable state sequence responsible for the first observations that have as its final state. The Viterbi path can be retrieved by saving back pointers that remember which state was used in the second equation. Let be the function that returns the value of used to compute if , or if . Then:

Here we’re using the standard definition of .

The complexity of this algorithm is .

## Example

Consider a village where all villagers are either healthy or have a fever and only the village doctor can determine whether each has a fever. The doctor diagnoses fever by asking patients how they feel. The villagers may only answer that they feel normal, dizzy, or cold.

The doctor believes that the health condition of his patients operate as a discrete . There are two states, “Healthy” and “Fever”, but the doctor cannot observe them directly; they are *hidden* from him. On each day, there is a certain chance that the patient will tell the doctor he/she is “normal”, “cold”, or “dizzy”, depending on their health condition.

The *observations* (normal, cold, dizzy) along with a *hidden* state (healthy, fever) form a hidden Markov model (HMM), and can be represented as follows in the Python programming language:

`obs = ('normal', 'cold', 'dizzy') states = ('Healthy', 'Fever') start_p = {'Healthy': 0.6, 'Fever': 0.4} trans_p = { 'Healthy' : {'Healthy': 0.7, 'Fever': 0.3}, 'Fever' : {'Healthy': 0.4, 'Fever': 0.6} } emit_p = { 'Healthy' : {'normal': 0.5, 'cold': 0.4, 'dizzy': 0.1}, 'Fever' : {'normal': 0.1, 'cold': 0.3, 'dizzy': 0.6} } `

In this piece of code, `start_probability`

represents the doctor’s belief about which state the HMM is in when the patient first visits (all he knows is that the patient tends to be healthy). The particular probability distribution used here is not the equilibrium one, which is (given the transition probabilities) approximately `{'Healthy': 0.57, 'Fever': 0.43}`

. The `transition_probability`

represents the change of the health condition in the underlying Markov chain. In this example, there is only a 30% chance that tomorrow the patient will have a fever if he is healthy today. The `emission_probability`

represents how likely the patient is to feel on each day. If he is healthy, there is a 50% chance that he feels normal; if he has a fever, there is a 60% chance that he feels dizzy.

The patient visits three days in a row and the doctor discovers that on the first day he feels normal, on the second day he feels cold, on the third day he feels dizzy. The doctor has a question: what is the most likely sequence of health conditions of the patient that would explain these observations? This is answered by the Viterbi algorithm.

` 1 def viterbi(obs, states, start_p, trans_p, emit_p): 2 V = [{}] 3 for st in states: 4 V[0][st] = {"prob": start_p[st] * emit_p[st][obs[0]], "prev": None} 5 # Run Viterbi when t > 0 6 for t in range(1, len(obs)): 7 V.append({}) 8 for st in states: 9 max_tr_prob = max(V[t-1][prev_st]["prob"]*trans_p[prev_st][st] for prev_st in states) 10 for prev_st in states: 11 if V[t-1][prev_st]["prob"] * trans_p[prev_st][st] == max_tr_prob: 12 max_prob = max_tr_prob * emit_p[st][obs[t]] 13 V[t][st] = {"prob": max_prob, "prev": prev_st} 14 break 15 for line in dptable(V): 16 print line 17 opt = [] 18 # The highest probability 19 max_prob = max(value["prob"] for value in V[-1].values()) 20 previous = None 21 # Get most probable state and its backtrack 22 for st, data in V[-1].items(): 23 if data["prob"] == max_prob: 24 opt.append(st) 25 previous = st 26 break 27 # Follow the backtrack till the first observation 28 for t in range(len(V) - 2, -1, -1): 29 opt.insert(0, V[t + 1][previous]["prev"]) 30 previous = V[t + 1][previous]["prev"] 31 32 print 'The steps of states are ' + ' '.join(opt) + ' with highest probability of %s' % max_prob 33 34 def dptable(V): 35 # Print a table of steps from dictionary 36 yield " ".join(("%12d" % i) for i in range(len(V))) 37 for state in V[0]: 38 yield "%.7s: " % state + " ".join("%.7s" % ("%f" % v[state]["prob"]) for v in V) `

The function `viterbi`

takes the following arguments: `obs`

is the sequence of observations, e.g. `['normal', 'cold', 'dizzy']`

; `states`

is the set of hidden states; `start_p`

is the start probability; `trans_p`

are the transition probabilities; and `emit_p`

are the emission probabilities. For simplicity of code, we assume that the observation sequence `obs`

is non-empty and that `trans_p[i][j]`

and `emit_p[i][j]`

is defined for all states i,j.

In the running example, the forward/Viterbi algorithm is used as follows:

`viterbi(obs, states, start_p, trans_p, emit_p) `

The output of the script is

`$ python viterbi_example.py 0 1 2 Healthy: 0.30000 0.08400 0.00588 Fever: 0.04000 0.02700 0.01512 The steps of states are Healthy Healthy Fever with highest probability of 0.01512 `

This reveals that the observations `['normal', 'cold', 'dizzy']`

were most likely generated by states `['Healthy', 'Healthy', 'Fever']`

. In other words, given the observed activities, the patient was most likely to have been healthy both on the first day when he felt normal as well as on the second day when he felt cold, and then he contracted a fever the third day.

The operation of Viterbi’s algorithm can be visualized by means of a . The Viterbi path is essentially the shortest path through this trellis. The trellis for the clinic example is shown below; the corresponding