Method get_state_sequence

hmm_class/hmmc.py:194–221

(self, x)

        return alpha[-1].sum()

    def get_state_sequence(self, x):
        # returns the most likely state sequence given observed sequence x
        # using the Viterbi algorithm
        T = len(x)

        # make the emission matrix B
        B = np.zeros((self.M, T))
        for j in range(self.M):
            for t in range(T):
                for k in range(self.K):
                    p = self.R[j,k] * mvn.pdf(x[t], self.mu[j,k], self.sigma[j,k])
                    B[j,t] += p

        # perform Viterbi as usual
        delta = np.zeros((T, self.M))
        psi = np.zeros((T, self.M))
        delta[0] = self.pi*B[:,0]
        for t in range(1, T):
            for j in range(self.M):
                delta[t,j] = np.max(delta[t-1]*self.A[:,j]) * B[j,t]
                psi[t,j] = np.argmax(delta[t-1]*self.A[:,j])

        # backtrack
        states = np.zeros(T, dtype=np.int32)
        states[T-1] = np.argmax(delta[T-1])
        for t in range(T-2, -1, -1):
            states[t] = psi[t+1, states[t+1]]
        return states

    def likelihood_multi(self, X):
        return np.array([self.likelihood(x) for x in X])