OLS missing data feature removal + docs update

HISTORY.rst

@@ -26,4 +26,9 @@ History
 0.2.0 (2024-03-2)
 ------------------
 
-* Overall project restructuring for optimisation
+* Overall project restructuring for optimisation
+
+0.2.1 (2024-03-3)
+------------------
+
+* Option to modify the OLS parameters used in the HLR

HLR/regression.py

@@ -18,7 +18,7 @@
 class HierarchicalLinearRegression:
     """Class for performing hierarchical linear regression analysis."""
 
-    def __init__(self, df, ivs_dict, dv, missing_data=None, ols_params=None):
+    def __init__(self, df, ivs_dict, dv, ols_params=None):
         """Initializes the HierarchicalLinearRegression class.
 
         Args:
@@ -44,11 +44,6 @@ def __init__(self, df, ivs_dict, dv, missing_data=None, ols_params=None):
             raise ValueError(f"{dv} is not a column in the DataFrame.")
         self.outcome_var = dv
 
-        if missing_data is not None:
-            if missing_data not in ['none', 'drop', 'raise']:
-                raise ValueError("missing_data must be either 'none', 'drop', or 'raise'.")
-        self.missing_data = missing_data if missing_data is not None else 'none'
-
         if ols_params is not None:
             if not isinstance(ols_params, dict):
                 raise ValueError("ols_params must be a dictionary.")
@@ -65,7 +60,7 @@ def fit_models(self):
             X = self.data[predictors]
             X_const = sm.add_constant(X)
             y = self.data[self.outcome_var]
-            model = sm.OLS(y, X_const, missing=self.missing_data).fit(**self.ols_params)
+            model = sm.OLS(y, X_const).fit(**self.ols_params)
             results[level] = model
         return results
 

README.md

@@ -150,11 +150,11 @@ This program is provided with no warranty of any kind and it is still under deve
 
 #### To-do
 Would be great if someone with more experience with packages would contribute with testing and the whole deployment process. Also, if someone would want to write documentation, that would be amazing.
-- docs
 - dict valus within df hard to read
 - ability to change OLS parameters
 - add t stats for coefficients
 - give option for output only some columns not all
+- add regression type option (eg, for logistic regression)
 
 #### Contributors
 [Toomas Erik Anijärv](https://github.com/teanijarv)

docs/usage.rst

@@ -2,38 +2,57 @@
 Usage
 =====
 
-See GitHub repository for the example dataset and Jupyter Notebook.
+Quick start guide to use hierarchical linear regression using HLR package.
 
-To use HLR - Hierarchical Linear Regression in a project::
+Initialising the HLR object
+---------------------------
 
-    import pandas as pd
-    from HLR import HierarchicalLinearRegression
+Let's first fetch some data and initiate the HLR object. We'll use the `penguins` dataset from `seaborn` for our example.
 
-    # Example dataframe which includes some columns which are also mentioned below
-    nba = pd.read_csv('example/NBA_train.csv')
+.. code-block:: python
 
-    # Define the models for hierarchical regression including predictors for each model
-    X = {1: ['PTS'], 
-         2: ['PTS', 'ORB'], 
-         3: ['PTS', 'ORB', 'BLK']}
+     import seaborn as sns
+     import pandas as pd
 
-    # Define the outcome variable
-    y = 'W'
+     # Load the example penguins dataset
+     df = sns.load_dataset('penguins')
+     df.dropna(inplace=True)
+     df = df[['bill_length_mm', 'bill_depth_mm', 'flipper_length_mm', 'body_mass_g']]
 
-    # Initiate the HLR object
-    hreg = HierarchicalLinearRegression(df, X, y)
+Initialising the HLR object and generating summary report
+---------------------------------------------------------
 
-    # Generate a summarised report as a dataframe which shows all linear regression models parameters and difference between the models
-    summary_report = hreg.summary()
-    display(summary_report)
+.. code-block:: python
 
-    # Run diagnostics on all the models (displayed output below only shows the first model)
-    hreg.diagnostics(verbose=True)
+     from HLR import HierarchicalLinearRegression
 
-    # Different plots
-    hreg.plot_studentized_residuals_vs_fitted()
-    hreg.plot_qq_residuals()
-    hreg.plot_influence()
-    hreg.plot_std_residuals()
-    hreg.plot_histogram_std_residuals()
-    hreg.plot_partial_regression()
+     # Define the independent variables for each model level
+     ivs_dict = {
+          1: ['bill_length_mm'],
+          2: ['bill_length_mm', 'bill_depth_mm'],
+          3: ['bill_length_mm', 'bill_depth_mm', 'flipper_length_mm']
+     }
+
+     # Define the dependent variable
+     dv = 'body_mass_g'
+
+     # Initialize the HierarchicalLinearRegression class
+     hlr = HierarchicalLinearRegression(df, ivs_dict, dv)
+     hlr.summary()
+
+Run diagnostics for testing assumptions
+---------------------------------------
+
+.. code-block:: python
+     diagnostics_dict = hlr.diagnostics(verbose=True)
+
+Plotting options for all model levels
+-------------------------------------
+
+.. code-block:: python
+     hlr.plot_studentized_residuals_vs_fitted()
+     hlr.plot_qq_residuals()
+     hlr.plot_influence()
+     hlr.plot_std_residuals()
+     hlr.plot_histogram_std_residuals()
+     hlr.plot_partial_regression()