1. Hidden Markov Models

2. A Hidden Markov Model consists of
A sequence of states {Xt|t  T} = {X1, X2, ... , XT} , and
A sequence of observations {Yt |t  T} = {Y1, Y2, ... , YT}

3. The sequence of states {X1, X2,
The sequence of states {X1, X2, ... , XT} form a Markov chain moving amongst the M states {1, 2, …, M}.
The observation Yt comes from a distribution that is determined by the current state of the process Xt. (or possibly past observations and past states).
The states, {X1, X2, ... , XT}, are unobserved (hence hidden).

4. The Hidden Markov ModelY1 Y2 Y3 YT … X1 X2 X3 XT
The Hidden Markov Model

5. Some basic problems:
from the observations {Y1, Y2, ... , YT}
1. Determine the sequence of states {X1, X2, ... , XT}.
2. Determine (or estimate) the parameters of the stochastic process that is generating the states and the observations.;

6. Examples

7. Example 1
A person is rolling two sets of dice (one is balanced, the other is unbalanced). He switches between the two sets of dice using a Markov transition matrix.
The states are the dice.
The observations are the numbers rolled each time.

8. Balanced Dice

9. Unbalanced Dice

10. Example 2 The Markov chain is two state.
The observations (given the states) are independent Normal.
Both mean and variance dependent on state.
HMM AR.xls

11. Example 3 –Dow Jones

12. Daily Changes Dow Jones

13. Hidden Markov Model??

14. Bear and Bull Market?

15. Speech Recognition
When a word is spoken the vocalization process goes through a sequence of states.
The sound produced is relatively constant when the process remains in the same state.
Recognizing the sequence of states and the duration of each state allows one to recognize the word being spoken.

16. The interval of time when the word is spoken is broken into small (possibly overlapping) subintervals.
In each subinterval one measures the amplitudes of various frequencies in the sound. (Using Fourier analysis). The vector of amplitudes Yt is assumed to have a multivariate normal distribution in each state with the mean vector and covariance matrix being state dependent.

17. Hidden Markov Models for Biological Sequence
Consider the Motif:
[AT][CG][AC][ACGT]*A[TG][GC]
Some realizations:
A C A A T G
T C A A C T A T C
A C A C - - A G C
A G A A T C
A C C G - - A T C

18. Hidden Markov model of the same motif : [AT][CG][AC][ACGT]*A[TG][GC].4
1.0 .6

19. Profile HMMs
Begin
End

20. Computing Likelihood
Let pij = P[Xt+1 = j|Xt = i] and P = (pij) = the MM transition matrix.
Let = P[X1 = i] and
= the initial distribution over the states.

21. P[Yt = yt |X1 = i1, X2 = i2, ... , Xt = it]
Now assume that
P[Yt = yt |X1 = i1, X2 = i2, ... , Xt = it]
= P[Yt = yt | Xt = it] = p(yt| ) =
Then
P[X1 = i1,X2 = i2..,XT = iT, Y1 = y1, Y2 = y2, ... , YT = yT]
= P[X = i, Y = y] =

22. Therefore
P[Y1 = y1, Y2 = y2, ... , YT = yT]
= P[Y = y]

23. In the case when Y1, Y2, ... , YT are continuous random variables or continuous random vectors, Let f(y| ) denote the conditional distribution of Yt given Xt = i. Then the joint density of Y1, Y2, ... , YT is given by
= f(y1, y2, ... , yT) = f(y)
where = f(yt| )

24. Efficient Methods for computing Likelihood
The Forward Method
Consider

27. The Backward Procedure

30. Prediction of states from the observations and the model:

31. The Viterbi Algorithm (Viterbi Paths)
Suppose that we know the parameters of the Hidden Markov Model.
Suppose in addition suppose that we have observed the sequence of observations Y1, Y2, ... , YT.
Now consider determining the sequence of States X1, X2, ... , XT.

32. Recall that
P[X1 = i1,... , XT = iT, Y1 = y1,... , YT = yT]
= P[X = i, Y = y] =
Consider the problem of determining the sequence of states, i1, i2, ... , iT , that maximizes the above probability.
This is equivalent to maximizing
P[X = i|Y = y] = P[X = i,Y = y] / P[Y = y]

33. The Viterbi Algorithm We want to maximize P[X = i, Y = y] =
Equivalently we want to minimize
U(i1, i2, ... , iT)
Where
= -ln (P[X = i, Y = y]) =

34. Minimization of U(i1, i2, ... , iT) can be achieved by Dynamic Programming.
This can be thought of as finding the shortest distance through the following grid of points.
By starting at the unique point in stage 0 and moving from a point in stage t to a point in stage t+1 in an optimal way. The distances between points in stage t and points in stage t+1 are equal to:

35. Dynamic Programming
Stage 0
Stage 1
Stage 2
Stage T-1
Stage T
...

36. By starting at the unique point in stage 0 and moving from a point in stage t to a point in stage t+1 in an optimal way.
The distances between points in stage t and points in stage t+1 are equal to:

37. Dynamic Programming
Stage 0
Stage 1
Stage 2
Stage T-1
Stage T
...

38. Dynamic Programming
...
Stage 0
Stage 1
Stage 2
Stage T-1
Stage T

39. Let
Then
i1 = 1, 2, …, M
and
it+1 = 1, 2, …, M; t = 1,…, T-2

40. Finally

41. Summary of calculations of Viterbi Path
i1 = 1, 2, …, M 2.
it+1 = 1, 2, …, M; t = 1,…, T-2 3.

42. An alternative approach to prediction of states from the observations and the model:
It can be shown that:

43. Forward Probabilities1. 2.

44. HMM generator (normal).xls
Backward Probabilities 1. 2.
HMM generator (normal).xls

45. Estimation of Parameters of a Hidden Markov Model
If both the sequence of observations Y1, Y2, ... , YT and the sequence of States X1, X2, ... , XT is observed Y1 = y1, Y2 = y2, ... , YT = yT, X1 = i1, X2 = i2, ... , XT = iT, then the Likelihood is given by:

46. the log-Likelihood is given by:

47. In this case the Maximum Likelihood estimates are:
= the MLE of qi computed from the observations yt where Xt = i.

48. MLE (states unknown)
If only the sequence of observations Y1 = y1, Y2 = y2, ... , YT = yT are observed then the Likelihood is given by:

49. It is difficult to find the Maximum Likelihood Estimates directly from the Likelihood function.
The Techniques that are used are
1. The Segmental K-means Algorithm
2. The Baum-Welch (E-M) Algorithm

50. The Segmental K-means Algorithm
In this method the parameters are adjusted to maximize
where is the Viterbi path

51. Consider this with the special case
Case: The observations {Y1, Y 2, ... , YT} are continuous Multivariate Normal with mean vector and covariance matrix when ,
i.e.

52. Pick arbitrarily M centroids a1, a2, … aM
Pick arbitrarily M centroids a1, a2, … aM. Assign each of the T observations yt (kT if multiple realizations are observed) to a state it by determining :
Then

53. And
Calculate the Viterbi path (i1, i2, …, iT) based on the parameters of step 2 and 3.
If there is a change in the sequence (i1, i2, …, iT) repeat steps 2 to 4.

54. The Baum-Welch (E-M) Algorithm
The E-M algorithm was designed originally to handle “Missing observations”.
In this case the missing observations are the states {X1, X2, ... , XT}.
Assuming a model, the states are estimated by finding their expected values under this model. (The E part of the E-M algorithm).

55. With these values the model is estimated by Maximum Likelihood Estimation (The M part of the E-M algorithm).
The process is repeated until the estimated model converges.

56. The E-M Algorithm Let denote the joint distribution of Y,X.
Consider the function:
Starting with an initial estimate of A sequence of estimates are formed by finding to maximize
with respect to .

57. The sequence of estimates
converge to a local maximum of the likelihood .

58. Example: Sampling from Mixtures
Let y1, y2, …, yn denote a sample from the density:
where
and

59. Suppose that m = 2 and let x1, x2, …, x1 denote independent random variables taking on the value 1 with probability f and 0 with probability 1- f.
Suppose that yi comes from the density
We will also assume that g(y|qi) is normal with mean miand standard deviation si.

60. Thus the joint distribution of x1, x2, …, xn and let y1, y2, …, yn is:

66. In the case of an HMM the log-Likelihood is given by:

67. Recall
and
Expected no. of transitions from
state i.

68. Expected no. of transitions from state i to
Let
Expected no. of transitions from state i to
state j.

69. The E-M Re-estimation Formulae
Case 1: The observations {Y1, Y2, ... , YT} are discrete with K possible values and

71. Case 2: The observations {Y1, Y 2,
Case 2: The observations {Y1, Y 2, ... , YT} are continuous Multivariate Normal with mean vector and covariance matrix when ,
i.e.

73. Measuring distance between two HMM’s
Let
and
denote the parameters of two different HMM models. We now consider defining a distance between these two models.

74. The Kullback-Leibler distance
Consider the two discrete distributions
and
( and in the continuous case)
then define

75. and in the continuous case:

76. These measures of distance between the two distributions are not symmetric but can be made symmetric by the following:

77. In the case of a Hidden Markov model.
where
The computation of in this case is formidable

78. Juang and Rabiner distance
Let denote a sequence of observations generated from the HMM with parameters:
Let
denote the optimal (Viterbi) sequence of states assuming HMM model

79. Then define:
and