v1.0.3

arma.egg-info/PKG-INFO

@@ -1,5 +1,5 @@
 Metadata-Version: 2.1
 Name: arma
-Version: 1.0.0
+Version: 1.0.3
 Summary:  ARMA, AR amd MA model process
 Author: Peter_Cheng

arma.egg-info/SOURCES.txt

@@ -1,3 +1,4 @@
+README.md
 setup.py
 arma/__init__.py
 arma/arma.py

arma/arma.py

@@ -4,6 +4,19 @@
 
 class ARMA:
     def __init__(self, p:int, q:int, phi:list, theta:list, c:float, mu:float, sigma:float):
+        '''
+        p: AR 的階次
+        q: MA 的階次
+        phi: AR 的係數
+        theta: MA 的係數
+        c: 常數項
+        --------------------
+        常態分佈的進行抽樣生成隨機誤差  N(mu, sigma)
+        mu: 隨機誤差的平均值
+        sigma: 隨機誤差的標準差
+        ex: ARMA(1, 1, [0.5], [0.5], 0, 0, 1)   
+        # ARMA(1, 1) process 係數設定 phi=0.5, theta=0.5, c=0, mu=0, sigma=1
+        '''
         if not isinstance(p, int) or not isinstance(q, int):
             raise ValueError("p and q must be integers")
         if len(phi) != p:
@@ -32,8 +45,12 @@ def plot(self):
     def save(self, file_name):
         return self._save(file_name)
 
+    def covariance(self,max_lag):
+        return self._covariance(max_lag)
+
+
+
 
-
     ##priveate function 
     def _simulate(self, num_samples):
         X = np.zeros(num_samples)
@@ -76,31 +93,100 @@ def _save(self, file_name):
             raise ValueError("file_name must end with .xlsx or .txt or .csv")
         return f"File saved as {file_name}"
 
+    def _covariance(arma, max_lag):
+        covariances = np.zeros(max_lag+1)
+        for k in range(max_lag+1):
+            if k > arma.q:
+                covariance = 0
+            else:
+                theta_padded = np.append(arma.theta, np.zeros(k))  # Pad theta with zeros for lag
+                covariance = arma.sigma**2 * np.sum(theta_padded[k:arma.q+k] * arma.theta[:arma.q])
+
+            # Add the contribution from the AR part
+            if k <= arma.p:
+                phi_padded = np.append(arma.phi, np.zeros(k))  # Pad phi with zeros for lag
+                covariance += np.sum(phi_padded[k:arma.p+k] * arma.phi[:arma.p])
+            covariances[k] = covariance
+            print(f'共變異數(lag {k}):{covariances[k]:.4f}')
+
+
+        return covariances
 
+
+####AR
 class AR(ARMA):
+        """
+        AR(p,phi, c, mu,sigma) process
+        p : AR 的階次   
+        phi : AR 的係數
+        c : 常數項
+        --------------------
+        常態分佈的進行抽樣生成隨機誤差  N(mu, sigma)
+        mu : 隨機誤差的平均值
+        sigma : 隨機誤差的標準差
+        ex: AR(1, [0.5], 0, 0, 1)
+        # AR(1) process 係數設定 phi=0.5, c=0, mu=0, sigma=1
+        """
         def __init__(self, p, phi, c, mu, sigma):
             super().__init__(p, 0, phi, [], c, mu, sigma)  # 調用父類的構造方法，MA部分設為空
         def simulate(self, num_samples=1):
             return self._simulate(num_samples)
         def statistics(self):
             return self._statistics()
+
+        # def plot(self):
+        #     return self._plot()
+
         def plot(self):
-            return self._plot()
+            plt.style.use('ggplot')
+            plt.figure(figsize=(10, 4))
+            plt.plot(self.memory, marker='o')
+            plt.title(f"AR({self.p}) Process")
+            plt.xlabel("Period")
+            plt.ylabel("Value")
+            plt.show()
+
         def save(self, file_name):
             return self._save(file_name)
+        def covariance(self,max_lag):
+            return self._covariance(max_lag)
+
 
 
-
-
+###MA
 class MA(ARMA):
+    """
+    MA(q,theta, c, mu,sigma) process
+    q : MA 的階次
+    theta : MA 的係數
+    c : 常數項
+    --------------------
+    常態分佈的進行抽樣生成隨機誤差  N(mu, sigma)
+    mu : 隨機誤差的平均值
+    sigma : 隨機誤差的標準差
+    ex: MA(1, [0.5], 0, 0, 1)
+    # MA(1) process 係數設定 theta=0.5, c=0, mu=0, sigma=1
+    """
     def __init__(self, q, theta, mu, c,sigma):
         super().__init__(0, q, [], theta, c, mu, sigma)
     def simulate(self, num_samples=1):
         return self._simulate(num_samples)
     def statistics(self):
         return self._statistics()
+
+    # def plot(self):
+    #     return self._plot()
+
     def plot(self):
-        return self._plot()
+        plt.style.use('ggplot')
+        plt.figure(figsize=(10, 4))
+        plt.plot(self.memory, marker='o')
+        plt.title(f"MA({self.q}) Process")
+        plt.xlabel("Period")
+        plt.ylabel("Value")
+        plt.show()
     def save(self, file_name):
         return self._save(file_name)
+    def covariance(self,max_lag):
+        return self._covariance(max_lag)
 

build/lib/arma/arma.py

@@ -4,6 +4,19 @@
 
 class ARMA:
     def __init__(self, p:int, q:int, phi:list, theta:list, c:float, mu:float, sigma:float):
+        '''
+        p: AR 的階次
+        q: MA 的階次
+        phi: AR 的係數
+        theta: MA 的係數
+        c: 常數項
+        --------------------
+        常態分佈的進行抽樣生成隨機誤差  N(mu, sigma)
+        mu: 隨機誤差的平均值
+        sigma: 隨機誤差的標準差
+        ex: ARMA(1, 1, [0.5], [0.5], 0, 0, 1)   
+        # ARMA(1, 1) process 係數設定 phi=0.5, theta=0.5, c=0, mu=0, sigma=1
+        '''
         if not isinstance(p, int) or not isinstance(q, int):
             raise ValueError("p and q must be integers")
         if len(phi) != p:
@@ -32,8 +45,12 @@ def plot(self):
     def save(self, file_name):
         return self._save(file_name)
 
+    def covariance(self,max_lag):
+        return self._covariance(max_lag)
+
+
+
 
-
     ##priveate function 
     def _simulate(self, num_samples):
         X = np.zeros(num_samples)
@@ -76,31 +93,100 @@ def _save(self, file_name):
             raise ValueError("file_name must end with .xlsx or .txt or .csv")
         return f"File saved as {file_name}"
 
+    def _covariance(arma, max_lag):
+        covariances = np.zeros(max_lag+1)
+        for k in range(max_lag+1):
+            if k > arma.q:
+                covariance = 0
+            else:
+                theta_padded = np.append(arma.theta, np.zeros(k))  # Pad theta with zeros for lag
+                covariance = arma.sigma**2 * np.sum(theta_padded[k:arma.q+k] * arma.theta[:arma.q])
+
+            # Add the contribution from the AR part
+            if k <= arma.p:
+                phi_padded = np.append(arma.phi, np.zeros(k))  # Pad phi with zeros for lag
+                covariance += np.sum(phi_padded[k:arma.p+k] * arma.phi[:arma.p])
+            covariances[k] = covariance
+            print(f'共變異數(lag {k}):{covariances[k]:.4f}')
+
+
+        return covariances
 
+
+####AR
 class AR(ARMA):
+        """
+        AR(p,phi, c, mu,sigma) process
+        p : AR 的階次   
+        phi : AR 的係數
+        c : 常數項
+        --------------------
+        常態分佈的進行抽樣生成隨機誤差  N(mu, sigma)
+        mu : 隨機誤差的平均值
+        sigma : 隨機誤差的標準差
+        ex: AR(1, [0.5], 0, 0, 1)
+        # AR(1) process 係數設定 phi=0.5, c=0, mu=0, sigma=1
+        """
         def __init__(self, p, phi, c, mu, sigma):
             super().__init__(p, 0, phi, [], c, mu, sigma)  # 調用父類的構造方法，MA部分設為空
         def simulate(self, num_samples=1):
             return self._simulate(num_samples)
         def statistics(self):
             return self._statistics()
+
+        # def plot(self):
+        #     return self._plot()
+
         def plot(self):
-            return self._plot()
+            plt.style.use('ggplot')
+            plt.figure(figsize=(10, 4))
+            plt.plot(self.memory, marker='o')
+            plt.title(f"AR({self.p}) Process")
+            plt.xlabel("Period")
+            plt.ylabel("Value")
+            plt.show()
+
         def save(self, file_name):
             return self._save(file_name)
+        def covariance(self,max_lag):
+            return self._covariance(max_lag)
+
 
 
-
-
+###MA
 class MA(ARMA):
+    """
+    MA(q,theta, c, mu,sigma) process
+    q : MA 的階次
+    theta : MA 的係數
+    c : 常數項
+    --------------------
+    常態分佈的進行抽樣生成隨機誤差  N(mu, sigma)
+    mu : 隨機誤差的平均值
+    sigma : 隨機誤差的標準差
+    ex: MA(1, [0.5], 0, 0, 1)
+    # MA(1) process 係數設定 theta=0.5, c=0, mu=0, sigma=1
+    """
     def __init__(self, q, theta, mu, c,sigma):
         super().__init__(0, q, [], theta, c, mu, sigma)
     def simulate(self, num_samples=1):
         return self._simulate(num_samples)
     def statistics(self):
         return self._statistics()
+
+    # def plot(self):
+    #     return self._plot()
+
     def plot(self):
-        return self._plot()
+        plt.style.use('ggplot')
+        plt.figure(figsize=(10, 4))
+        plt.plot(self.memory, marker='o')
+        plt.title(f"MA({self.q}) Process")
+        plt.xlabel("Period")
+        plt.ylabel("Value")
+        plt.show()
     def save(self, file_name):
         return self._save(file_name)
+    def covariance(self,max_lag):
+        return self._covariance(max_lag)
 

file_name=arma_model.csv

@@ -0,0 +1,22 @@
+0
+3.1899210099327746
+3.789687398455347
+3.562788655308019
+2.7198282727217
+2.2734373599966227
+3.314763502530126
+3.3618291920854193
+2.411443170352059
+2.0569450953029156
+0.6436660274221947
+3.3751642110778004
+2.621653486511918
+1.8234711457067176
+2.6057025080104284
+3.0562586409286823
+2.346018988966077
+2.198230516152887
+1.9029125879564355
+1.3133840074780667
+0.09505671985988684
+0.0

setup.py

@@ -2,7 +2,7 @@
 
 setup(
     name='arma',
-    version='1.0.2',
+    version='1.0.3',
     description=' ARMA, AR amd MA model process',
     author='Peter_Cheng',
     packages=find_packages(),