Update manual.rst

docs/manual.rst

@@ -1 +1,162 @@
+Manual pySTARMA library
+================
+This file contains the manual for the usage of the pySTARMA library
 
+STARMA object
+-----------------
+
+Description
+~~~~~~~~~~~~~~~~~~~~~~
+The STARMA class can be used to estimate STARMA models. The method ``STARMA.fit()`` performs the estimation of the parameters. The method ``STARMA.predict()`` executes the forecast (still in the development stage). The method ``STARMA.get_model()`` returns the full model. The ``STARMA.get_item()`` method returns a selected property of the model (`see Return Values STARMA`_).
+
+Usage
+~~~~~~~~~~~~~~~~~~~~~~
+.. code-block:: python
+  
+  model = sm.STARMA(p, q, ts_matrix, wa_matrices, iterations(optional))
+  model.fit()
+  model.get_model()
+  model.get_item()
+  
+Example  
+~~~~~~~~~~~~~~~~~~~~~~
+.. code-block:: python
+
+  from pySTARMA import starma_model as sm
+  
+  #Create instance of STARMA
+  model = sm.STARMA(5, 2, time_series, wa_matrices, 3)
+  
+  #Estimate parameters
+  model.fit()
+
+  #Print explicit item 
+  print model.get_item('bic')
+  
+Attributes
+~~~~~~~~~~~~~~~~~~~~~~
++---------------------+---------------------------------------------+
+| Attribute           | Value                                       |
++=====================+=============================================+
+|p                    |Number or list of autoregressive parameters  |
++---------------------+---------------------------------------------+
+|q                    | Number or list of moving average parameters |
++---------------------+---------------------------------------------+
+|ts_matrix            | Time series matrix                          |
++---------------------+---------------------------------------------+
+|wa_matrices          | List of adjacency matrices                  |
++---------------------+---------------------------------------------+
+|iterations(optional) | Number of iteration of kalman filtering,    |
+|                     | only for estimation of moving average       |
+|                     | parameters                                  |
++---------------------+---------------------------------------------+
+
+Returnvalues
+~~~~~~~~~~~~~~~~~~~~~~
+
+.. _`see Return Values STARMA`:
+
+A dictionary is returned as a 'model' with the following values:
+
++---------------------+---------------------------------------------+
+| Value               | Description                                 |
++=====================+=============================================+
+|residuals            | Matrix with estimated residuals             |
++---------------------+---------------------------------------------+
+|phi                  | Matrix with estimated AR-parameters         |
++---------------------+---------------------------------------------+
+|phi_tvalue           | Matrix with estimated AR-t-values           |
++---------------------+---------------------------------------------+
+|phi_pvalue           | Matrix with estimated AR-p-values           |
++---------------------+---------------------------------------------+
+|theta                | Matrix with estimated MA-parameters         |
++---------------------+---------------------------------------------+
+|theta_tvalue         | Matrix with estimated MA-t-values           |
++---------------------+---------------------------------------------+
+|theta_pvalue         | Matrix with estimated MA-p-values           |
++---------------------+---------------------------------------------+
+|sigma2               | Standard deviation                          |
++---------------------+---------------------------------------------+
+|bic                  | Bayessche informationcriterion              |
++---------------------+---------------------------------------------+
+
+STARIMA object
+-----------------
+
+Description
+~~~~~~~~~~~~~~~~~~~~~~
+The STARIMA class can be used to estimate STARIMA models. The method ``STARIMA.fit()`` performs the estimation of the parameters. The method ``STARIMA.predict()`` executes the forecast (still in the development stage). The method ``STARIMA.get_model()`` returns the full model. The ``STARIMA.get_item()`` method returns a selected property of the model (`see Return Values STARIMA`_).
+
+Usage
+~~~~~~~~~~~~~~~~~~~~~~
+.. code-block:: python
+  
+  model = sm.STARIMA(p, q, d, ts_matrix, wa_matrices, iterations(optional))
+  model.fit()
+  model.get_model()
+  model.get_item()
+  
+Example  
+~~~~~~~~~~~~~~~~~~~~~~
+.. code-block:: python
+
+  from pySTARMA import starma_model as sm
+  
+  #Create instance of STARIMA
+  model = sm.STARMA(5, 2, (1,), time_series, wa_matrices, 3)
+  
+  #Estimate parameters
+  model.fit()
+
+  #Print explicit item 
+  print model.get_item('bic')
+  
+Attributes
+~~~~~~~~~~~~~~~~~~~~~~
++---------------------+---------------------------------------------+
+| Attribute           | Value                                       |
++=====================+=============================================+
+|p                    |Number or list of autoregressive parameters  |
++---------------------+---------------------------------------------+
+|q                    | Number or list of moving average parameters |
++---------------------+---------------------------------------------+
+|d                    | List of numbers of differentiatio           |
++---------------------+---------------------------------------------+
+|ts_matrix            | Time series matrix                          |
++---------------------+---------------------------------------------+
+|wa_matrices          | List of adjacency matrices                  |
++---------------------+---------------------------------------------+
+|iterations(optional) | Number of iteration of kalman filtering,    |
+|                     | only for estimation of moving average       |
+|                     | parameters                                  |
++---------------------+---------------------------------------------+
+
+Return Values
+~~~~~~~~~~~~~~~~~~~~~~
+
+.. _`see Returnvalues STARIMA`:
+
+A dictionary is returned as a 'model' with the following values:
+
+
++---------------------+---------------------------------------------+
+| Value               | Description                                 |
++=====================+=============================================+
+|residuals            | Matrix with estimated residuals             |
++---------------------+---------------------------------------------+
+|phi                  | Matrix with estimated AR-parameters         |
++---------------------+---------------------------------------------+
+|phi_tvalue           | Matrix with estimated AR-t-values           |
++---------------------+---------------------------------------------+
+|phi_pvalue           | Matrix with estimated AR-p-values           |
++---------------------+---------------------------------------------+
+|theta                | Matrix with estimated MA-parameters         |
++---------------------+---------------------------------------------+
+|theta_tvalue         | Matrix with estimated MA-t-values           |
++---------------------+---------------------------------------------+
+|theta_pvalue         | Matrix with estimated MA-p-values           |
++---------------------+---------------------------------------------+
+|sigma2               | Standard deviation                          |
++---------------------+---------------------------------------------+
+|bic                  | Bayessche informationcriterion              |
++---------------------+---------------------------------------------+