Scribd
Upload
24 views2 pages

HMM Forward Backward Algorithm

The document describes the Forward and Backward algorithms used in a Hidden Markov Model (HMM) with two states (Rainy and Sunny) and two observations (Walk and Shop). It details the initialization, induction, and termination steps for both algorithms, ultimately calculating the probability of the observation sequence as P(O|lambda) = 0.03324. The results from both algorithms confirm the same probability value.

Uploaded by

sriranganvel2004
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
24 views2 pages

HMM Forward Backward Algorithm

The document describes the Forward and Backward algorithms used in a Hidden Markov Model (HMM) with two states (Rainy and Sunny) and two observations (Walk and Shop). It details the initialization, induction, and termination steps for both algorithms, ultimately calculating the probability of the observation sequence as P(O|lambda) = 0.03324. The results from both algorithms confirm the same probability value.

Uploaded by

sriranganvel2004
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Forward and Backward Algorithm in Hidden Markov Model (HMM)

Model Parameters:

- States (N = 2): Rainy (S1), Sunny (S2)

- Observations (M = 2): Walk, Shop

- Observation sequence (T = 3): O = [Walk, Shop, Walk]

HMM Parameters:

- Initial Probabilities (pi): pi = [0.6, 0.4]

- Transition Probabilities (A):

A = [[0.7, 0.3],

[0.4, 0.6]]

- Emission Probabilities (B):

B = [[P(Walk|Rainy) = 0.1, P(Shop|Rainy) = 0.4],

[P(Walk|Sunny) = 0.6, P(Shop|Sunny) = 0.3]]

Forward Algorithm:

1. Initialization (t=1):

alpha1(1) = 0.6 * 0.1 = 0.06

alpha1(2) = 0.4 * 0.6 = 0.24

2. Induction (t=2, observation=Shop):

alpha2(1) = (0.06*0.7 + 0.24*0.4) * 0.4 = 0.0552

alpha2(2) = (0.06*0.3 + 0.24*0.6) * 0.3 = 0.0486

3. Induction (t=3, observation=Walk):

alpha3(1) = (0.0552*0.7 + 0.0486*0.4) * 0.1 = 0.005808


alpha3(2) = (0.0552*0.3 + 0.0486*0.6) * 0.6 = 0.027432

4. Termination:

P(O|lambda) = alpha3(1) + alpha3(2) = 0.03324

Backward Algorithm:

1. Initialization (t=3):

beta3(1) = beta3(2) = 1

2. Induction (t=2):

beta2(1) = 0.7*0.1 + 0.3*0.6 = 0.25

beta2(2) = 0.4*0.1 + 0.6*0.6 = 0.4

3. Induction (t=1):

beta1(1) = 0.7*0.4*0.25 + 0.3*0.3*0.4 = 0.106

beta1(2) = 0.4*0.4*0.25 + 0.6*0.3*0.4 = 0.112

4. Termination:

P(O|lambda) = 0.6*0.1*0.106 + 0.4*0.6*0.112 = 0.03324

Final Result:

P(O|lambda) = 0.03324 (using both forward and backward algorithms)

You might also like