将数据拟合到hmm.MultinomialHMM

Question

我试图预测使用hmmlearn库给出一些数据的最佳序列，但是我收到一个错误。我的代码是：

from hmmlearn import hmm 
trans_mat = np.array([[0.2,0.6,0.2],[0.4,0.0,0.6],[0.1,0.2,0.7]]) 
emm_mat = np.array([[0.2,0.1,0.1,0.1,0.1,0.1,0.1,0.1,0.1],[0.1,0.1,0.1,0.1,0.2,0.1,0.1,0.1,0.1],[0.1,0.1,0.1,0.1,0.1,0.1,0.1,0.1,0.2]]) 
start_prob = np.array([0.3,0.4,0.3]) 
X = [3,4,5,6,7] 
model = GaussianHMM(n_components = 3, n_iter = 1000) 
X = np.array(X) 
model.startprob_ = start_prob 
model.transmat_ = trans_mat 
model.emissionprob_ = emm_mat 

# Predict the optimal sequence of internal hidden state 
x = model.fit([X]) 

print(model.decode([X])) 

，但我得到一个错误说：

Traceback (most recent call last): 
    File "hmm_loyalty.py", line 55, in <module> 
    x = model.fit([X]) 
    File "build/bdist.macosx-10.6-x86_64/egg/hmmlearn/base.py", line 421, in fit 
    File "build/bdist.macosx-10.6-x86_64/egg/hmmlearn/hmm.py", line 183, in _init 
    File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/sklearn/cluster/k_means_.py", line 785, in fit 
    X = self._check_fit_data(X) 
    File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/sklearn/cluster/k_means_.py", line 758, in _check_fit_data 
X.shape[0], self.n_clusters)) 
ValueError: n_samples=1 should be >= n_clusters=3 

任何人有任何想法，这是什么意思和我能做些什么来解决它？

Answer 1

是有一些问题与您的代码：

1. model是一个GaussianHMM。你可能想要MultinomialHMM。
2. 输入X的形状不正确。对于MultinomialHMM X必须具有形状(n_samples, 1)，因为观察结果是一维的。
3. 你不想要fit，除非需要估计一些模型参数，这里不是这种情况。

这里有一个工作版本

import numpy as np 
from hmmlearn import hmm 

model = hmm.MultinomialHMM(n_components=3) 
model.startprob_ = np.array([0.3, 0.4, 0.3]) 
model.transmat_ = np.array([[0.2, 0.6, 0.2], 
          [0.4, 0.0, 0.6], 
          [0.1, 0.2, 0.7]]) 
model.emissionprob_ = np.array([[0.2, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1], 
           [0.1, 0.1, 0.1, 0.1, 0.2, 0.1, 0.1, 0.1, 0.1], 
           [0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.2]]) 

# Predict the optimal sequence of internal hidden state 
X = np.atleast_2d([3, 4, 5, 6, 7]).T 
print(model.decode(X))