Release 0.2.0

README.md

@@ -23,7 +23,7 @@ For more details, see our [website](https://yuta-nakahara.github.io/BayesML/ "Ba
 
 Please use the following commands to install BayesML.
 
-``` shell
+``` bash
 pip install bayesml
 ```
 
@@ -33,6 +33,7 @@ The following are required.
 * NumPy (>= 1.20)
 * SciPy (>= 1.7)
 * MatplotLib (>= 3.5)
+* Scikit-learn (>= 1.1)
 
 ## Example
 
@@ -53,11 +54,11 @@ gen_model.visualize_model()
 ```
 
 >theta:0.7  
->x0:[1 1 1 1 1 0 1 0 0 1 1 1 1 0 1 1 0 1 1 1]  
->x1:[1 0 1 1 1 1 1 1 0 0 1 1 1 1 1 1 1 1 1 1]  
->x2:[0 0 1 1 0 1 0 1 1 1 1 1 1 0 1 0 1 1 1 1]  
->x3:[1 0 1 1 1 1 1 0 0 0 1 0 0 1 0 1 1 0 1 0]  
->x4:[1 1 0 1 0 1 1 1 0 1 1 1 0 0 1 1 1 1 1 1]  
+>x0:[1 1 1 0 1 1 1 0 1 1 1 1 1 1 0 1 1 1 0 1]  
+>x1:[1 1 0 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0]  
+>x2:[1 0 1 1 0 1 1 1 0 1 1 1 1 1 0 0 1 1 1 1]  
+>x3:[1 1 1 0 1 1 0 1 0 0 0 0 1 0 1 1 1 1 1 1]  
+>x4:[0 0 1 0 0 0 1 1 1 1 1 1 1 1 0 0 1 1 1 1]  
 >![bernoulli_example1](./doc/images/README_ex_img1.png)
 
 After confirming that the frequency of occurrence of 1 is around `theta=0.7`, we generate a sample and store it to variable `x`.
@@ -99,9 +100,9 @@ print(learn_model.estimate_params(loss='abs'))
 print(learn_model.estimate_params(loss='0-1'))
 ```
 
->0.6428571428571429  
->0.6474720009710451  
->0.6578947368421053  
+>0.7380952380952381  
+>0.7457656349087012  
+>0.7631578947368421  
 
 Different settings of a loss function yield different optimal estimates.
 
@@ -115,8 +116,12 @@ The following packages are currently available. In this library, a probabilistic
 * [Normal model](https://yuta-nakahara.github.io/BayesML/bayesml.normal.html "BayesML Normal Model")
 * [Multivariate normal model](https://yuta-nakahara.github.io/BayesML/bayesml.multivariate_normal.html "BayesML Multivariate Normal Model")
 * [Exponential model](https://yuta-nakahara.github.io/BayesML/bayesml.exponential.html "BayesML Exponential Model")
+* [Gaussian mixture model](https://yuta-nakahara.github.io/BayesML/bayesml.gaussianmixture.html "BayesML Gaussian Mixture Model")
 * [Linear regression model](https://yuta-nakahara.github.io/BayesML/bayesml.linearregression.html "BayesML Lenear Regression Model")
+* [Meta-tree model](https://yuta-nakahara.github.io/BayesML/bayesml.metatree.html "BayesML Meta-tree Model")
 * [Autoregressive model](https://yuta-nakahara.github.io/BayesML/bayesml.autoregressive.html "BayesML Autoregressive Model")
+* [Hidden Markov normal model](https://yuta-nakahara.github.io/BayesML/bayesml.hiddenmarkovnormal.html "BayesML Hidden Markov Normal Model")
+* [Context tree model](https://yuta-nakahara.github.io/BayesML/bayesml.contexttree.html "BayesML Context Tree Model")
 
 In the future, we will add packages to deal with a mixture normal model and a hidden Markov model, which are difficult to perform exact Bayesian inference, by using variational Bayes methods.
 
@@ -131,11 +136,8 @@ When you use BayesML for your academic work, please provide the following biblio
 Plain text
 
 ```
-Y. Nakahara, N. Ichijo, K. Shimada, 
-K. Tajima, K. Horinouchi, L. Ruan, 
-N. Namegaya, R. Maniwa, T. Ishiwatari, 
-W. Yu, Y. Iikubo, S. Saito, 
-K. Kazama, T. Matsushima, ``BayesML,'' 
+Y. Nakahara, N. Ichijo, K. Shimada, Y. Iikubo, 
+S. Saito, K. Kazama, T. Matsushima, ``BayesML 0.2.0,'' 
 [Online] https://github.com/yuta-nakahara/BayesML
 ```
 
@@ -144,11 +146,9 @@ BibTeX
 ``` bibtex
 @misc{bayesml,
   author = {Nakahara Yuta and Ichijo Naoki and Shimada Koshi and
-            Tajima Keito and Horinouchi Kohei and Ruan Luyu and
-            Namegaya Noboru and Maniwa Ryota and Ishiwatari Taisuke and
-            Yu Wenbin and Iikubo Yuji and Saito Shota and Kazama Koki and
-            Matsushima Toshiyasu},
-  title = {BayesML},
+            Iikubo Yuji and Saito Shota and Kazama Koki and
+            Matsushima Toshiyasu}
+  title = {BayesML 0.2.0},
   howpublished = {\url{https://github.com/yuta-nakahara/BayesML}},
   year = {2022}
 }

README_jp.md

@@ -22,7 +22,7 @@ BayesMLは以下の特徴を持っています．
 
 以下のコマンドによりインストール可能です．
 
-``` shell
+``` bash
 pip install bayesml
 ```
 
@@ -32,6 +32,7 @@ BayesMLの実行には以下が必要です．
 * NumPy (>= 1.20)
 * SciPy (>= 1.7)
 * MatplotLib (>= 3.5)
+* Scikit-learn (>= 1.1)
 
 ## 実行例
 
@@ -52,11 +53,11 @@ gen_model.visualize_model()
 ```
 
 >theta:0.7  
->x0:[1 1 1 1 1 0 1 0 0 1 1 1 1 0 1 1 0 1 1 1]  
->x1:[1 0 1 1 1 1 1 1 0 0 1 1 1 1 1 1 1 1 1 1]  
->x2:[0 0 1 1 0 1 0 1 1 1 1 1 1 0 1 0 1 1 1 1]  
->x3:[1 0 1 1 1 1 1 0 0 0 1 0 0 1 0 1 1 0 1 0]  
->x4:[1 1 0 1 0 1 1 1 0 1 1 1 0 0 1 1 1 1 1 1]  
+>x0:[1 1 1 0 1 1 1 0 1 1 1 1 1 1 0 1 1 1 0 1]  
+>x1:[1 1 0 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0]  
+>x2:[1 0 1 1 0 1 1 1 0 1 1 1 1 1 0 0 1 1 1 1]  
+>x3:[1 1 1 0 1 1 0 1 0 0 0 0 1 0 1 1 1 1 1 1]  
+>x4:[0 0 1 0 0 0 1 1 1 1 1 1 1 1 0 0 1 1 1 1]  
 >![bernoulli_example1](./doc/images/README_ex_img1.png)
 
 1の出現頻度が`theta=0.7`程度であることを確認したら，サンプルを生成し変数`x`に保存します．
@@ -96,9 +97,9 @@ print(learn_model.estimate_params(loss='abs'))
 print(learn_model.estimate_params(loss='0-1'))
 ```
 
->0.6428571428571429  
->0.6474720009710451  
->0.6578947368421053  
+>0.7380952380952381  
+>0.7457656349087012  
+>0.7631578947368421  
 
 損失関数の設定が異なると，そのもとでの最適な推定値も異なることがわかります．
 
@@ -112,8 +113,12 @@ print(learn_model.estimate_params(loss='0-1'))
 * [正規モデル](https://yuta-nakahara.github.io/BayesML/bayesml.normal.html "BayesML Normal Model")
 * [多変量正規モデル](https://yuta-nakahara.github.io/BayesML/bayesml.multivariate_normal.html "BayesML Multivariate Normal Model")
 * [指数モデル](https://yuta-nakahara.github.io/BayesML/bayesml.exponential.html "BayesML Exponential Model")
+* [混合正規モデル](https://yuta-nakahara.github.io/BayesML/bayesml.gaussianmixture.html "BayesML Gaussian Mixture Model")
 * [線形回帰モデル](https://yuta-nakahara.github.io/BayesML/bayesml.linearregression.html "BayesML Lenear Regression Model")
+* [メタツリーモデル](https://yuta-nakahara.github.io/BayesML/bayesml.metatree.html "BayesML Meta-tree Model")
 * [自己回帰モデル](https://yuta-nakahara.github.io/BayesML/bayesml.autoregressive.html "BayesML Autoregressive Model")
+* [隠れマルコフモデル](https://yuta-nakahara.github.io/BayesML/bayesml.hiddenmarkovnormal.html "BayesML Hidden Markov Normal Model")
+* [文脈木モデル](https://yuta-nakahara.github.io/BayesML/bayesml.contexttree.html "BayesML Context Tree Model")
 
 また，今後は混合正規モデルや隠れマルコフモデルなどの厳密なベイズ推論が困難なモデルを変分ベイズ法で学習するパッケージが追加される予定です．
 
@@ -128,11 +133,8 @@ BayesMLへのコントリビューションを考えてくださってありが
 プレーンテキスト
 
 ```
-Y. Nakahara, N. Ichijo, K. Shimada, 
-K. Tajima, K. Horinouchi, L. Ruan, 
-N. Namegaya, R. Maniwa, T. Ishiwatari, 
-W. Yu, Y. Iikubo, S. Saito, 
-K. Kazama, T. Matsushima, ``BayesML,'' 
+Y. Nakahara, N. Ichijo, K. Shimada, Y. Iikubo, 
+S. Saito, K. Kazama, T. Matsushima, ``BayesML,'' 
 [Online] https://github.com/yuta-nakahara/BayesML
 ```
 
@@ -141,9 +143,7 @@ BibTeX
 ``` bibtex
 @misc{bayesml,
   author = {Nakahara Yuta and Ichijo Naoki and Shimada Koshi and
-            Tajima Keito and Horinouchi Kohei and Ruan Luyu and
-            Namegaya Noboru and Maniwa Ryota and Ishiwatari Taisuke and
-            Yu Wenbin and Iikubo Yuji and Saito Shota and Kazama Koki and
+            Iikubo Yuji and Saito Shota and Kazama Koki and
             Matsushima Toshiyasu}
   title = {BayesML},
   howpublished = {\url{https://github.com/yuta-nakahara/BayesML}},

bayesml/__init__.py

@@ -6,6 +6,10 @@
 from . import multivariate_normal
 from . import normal
 from . import poisson
+from . import metatree
+from . import contexttree
+from . import gaussianmixture
+from . import hiddenmarkovnormal
 
 __all__ = ['bernoulli',
            'categorical',
@@ -14,5 +18,9 @@
            'linearregression',
            'multivariate_normal',
            'normal',
-           'poisson'
+           'poisson',
+           'metatree',
+           'contexttree',
+           'gaussianmixture',
+           'hiddenmarkovnormal',
            ]

bayesml/_check.py

@@ -1,6 +1,8 @@
 # Code Author
 # Yuta Nakahara <yuta.nakahara@aoni.waseda.jp>
 # Yuji Iikubo <yuji-iikubo.8@fuji.waseda.jp>
+# Yasushi Esaki <esakiful@gmail.com>
+# Jun Nishikawa <jun.b.nishikawa@gmail.com>
 import numpy as np
 
 _EPSILON = np.sqrt(np.finfo(np.float64).eps)
@@ -45,12 +47,31 @@ def nonneg_ints(val,val_name,exception_class):
             return val
     raise(exception_class(val_name + " must be int or a numpy.ndarray whose dtype is int. Its values must be non-negative (including 0)."))
 
+def int_vec(val,val_name,exception_class):
+    if type(val) is np.ndarray:
+        if np.issubdtype(val.dtype,np.integer) and val.ndim == 1:
+            return val
+    raise(exception_class(val_name + " must be a 1-dimensional numpy.ndarray whose dtype is int."))
+
 def nonneg_int_vec(val,val_name,exception_class):
     if type(val) is np.ndarray:
         if np.issubdtype(val.dtype,np.integer) and val.ndim == 1 and np.all(val>=0):
             return val
     raise(exception_class(val_name + " must be a 1-dimensional numpy.ndarray whose dtype is int. Its values must be non-negative (including 0)."))
 
+def nonneg_int_vecs(val,val_name,exception_class):
+    if type(val) is np.ndarray:
+        if np.issubdtype(val.dtype,np.integer) and val.ndim >= 1 and np.all(val>=0):
+            return val
+    raise(exception_class(val_name + " must be a numpy.ndarray whose ndim >= 1 and dtype is int. Its values must be non-negative (including 0)."))
+
+
+def nonneg_float_vec(val,val_name,exception_class):
+    if type(val) is np.ndarray:
+        if np.issubdtype(val.dtype,np.floating) and val.ndim == 1 and np.all(val>=0):
+            return val
+    raise(exception_class(val_name + " must be a 1-dimensional numpy.ndarray whose dtype is float. Its values must be non-negative (including 0)."))
+
 def int_of_01(val,val_name,exception_class):
     if np.issubdtype(type(val),np.integer):
         if val == 0 or val ==1:
@@ -100,6 +121,22 @@ def pos_def_sym_mat(val,val_name,exception_class):
         pass
     raise(exception_class(val_name + " must be a positive definite symmetric 2-dimensional numpy.ndarray."))
 
+def sym_mats(val,val_name,exception_class):
+    if type(val) is np.ndarray:
+        if val.ndim >= 2 and val.shape[-1] == val.shape[-2]:
+            if np.allclose(val, np.swapaxes(val,-1,-2)):
+                return val
+    raise(exception_class(val_name + " must be a symmetric 2-dimensional numpy.ndarray."))
+
+def pos_def_sym_mats(val,val_name,exception_class):
+    sym_mats(val,val_name,exception_class)
+    try:
+        np.linalg.cholesky(val)
+        return val
+    except np.linalg.LinAlgError:
+        pass
+    raise(exception_class(val_name + " must be a positive definite symmetric 2-dimensional numpy.ndarray."))
+
 def float_(val,val_name,exception_class):
     if np.issubdtype(type(val),np.floating):
         return val
@@ -155,6 +192,14 @@ def float_vecs(val,val_name,exception_class):
             return val
     raise(exception_class(val_name + " must be a numpy.ndarray whose ndim >= 1."))
 
+def pos_float_vecs(val,val_name,exception_class):
+    if type(val) is np.ndarray:
+        if np.issubdtype(val.dtype,np.integer) and val.ndim >= 1 and np.all(val>0):
+            return val.astype(float)
+        if np.issubdtype(val.dtype,np.floating) and val.ndim >= 1 and np.all(val>0.0):
+            return val
+    raise(exception_class(val_name + " must be a 1-dimensional numpy.ndarray. Its values must be positive (not including 0)"))
+
 def float_vec_sum_1(val,val_name,exception_class):
     if type(val) is np.ndarray:
         if np.issubdtype(val.dtype,np.integer) and val.ndim == 1 and abs(val.sum() - 1.) <= _EPSILON:
@@ -163,6 +208,14 @@ def float_vec_sum_1(val,val_name,exception_class):
             return val
     raise(exception_class(val_name + " must be a 1-dimensional numpy.ndarray, and the sum of its elements must equal to 1."))
 
+def float_vecs_sum_1(val,val_name,exception_class):
+    if type(val) is np.ndarray:
+        if np.issubdtype(val.dtype,np.integer) and val.ndim >= 1 and np.all(np.abs(np.sum(val, axis=-1) - 1.) <= _EPSILON):
+            return val.astype(float)
+        if np.issubdtype(val.dtype,np.floating) and val.ndim >= 1 and np.all(np.abs(np.sum(val, axis=-1) - 1.) <= _EPSILON):
+            return val
+    raise(exception_class(val_name + " must be a numpy.ndarray whose ndim >= 1, and the sum along the last dimension must equal to 1."))
+
 def int_(val,val_name,exception_class):   
     if np.issubdtype(type(val),np.integer):
         return val
@@ -189,3 +242,15 @@ def onehot_vecs(val,val_name,exception_class):
         if np.issubdtype(val.dtype,np.integer) and val.ndim >= 1 and np.all(val >= 0) and np.all(val.sum(axis=-1)==1):
             return val
     raise(exception_class(val_name + " must be a numpy.ndarray whose dtype is int and whose last axis constitutes one-hot vectors."))
+
+def int_vecs(val,val_name,exception_class):
+    if type(val) is np.ndarray:
+        if np.issubdtype(val.dtype,np.integer) and val.ndim >= 1:
+            return val
+    raise(exception_class(val_name + " must be a numpy.ndarray whose dtype is int and ndim >= 1."))
+
+def shape_consistency(val: int, val_name: str, correct: int, correct_name: str, exception_class):
+    if val != correct:
+        message = (f"{val_name} must coincide with {correct_name}: "
+                   + f"{val_name} = {val}, {correct_name} = {correct}")
+        raise(exception_class(message))