Update app.py

app.py

@@ -37,6 +37,15 @@ class RSM_BoxBehnken:
         Inicializa la clase con los datos del diseño Box-Behnken.
         """
         self.data = data.copy()
+        # Crear columnas cuadráticas con nombres consistentes
+        self.data[f'{x1_name}_sq'] = self.data[x1_name] ** 2
+        self.data[f'{x2_name}_sq'] = self.data[x2_name] ** 2
+        self.data[f'{x3_name}_sq'] = self.data[x3_name] ** 2
+        # Crear columnas de interacción
+        self.data[f'{x1_name}_{x2_name}'] = self.data[x1_name] * self.data[x2_name]
+        self.data[f'{x1_name}_{x3_name}'] = self.data[x1_name] * self.data[x3_name]
+        self.data[f'{x2_name}_{x3_name}'] = self.data[x2_name] * self.data[x3_name]
+        
         self.model = None
         self.model_simplified = None
         self.optimized_results = None
@@ -94,8 +103,8 @@ class RSM_BoxBehnken:
         Ajusta el modelo de segundo orden completo a los datos.
         """
         formula = f'{self.y_name} ~ {self.x1_name} + {self.x2_name} + {self.x3_name} + ' \
-                  f'I({self.x1_name}**2) + I({self.x2_name}**2) + I({self.x3_name}**2) + ' \
-                  f'{self.x1_name}:{self.x2_name} + {self.x1_name}:{self.x3_name} + {self.x2_name}:{self.x3_name}'
+                  f'{self.x1_name}_sq + {self.x2_name}_sq + {self.x3_name}_sq + ' \
+                  f'{self.x1_name}_{self.x2_name} + {self.x1_name}_{self.x3_name} + {self.x2_name}_{self.x3_name}'
         self.model = smf.ols(formula, data=self.data).fit()
         print("Modelo Completo:")
         print(self.model.summary())
@@ -106,7 +115,7 @@ class RSM_BoxBehnken:
         Ajusta el modelo de segundo orden a los datos, eliminando términos no significativos.
         """
         formula = f'{self.y_name} ~ {self.x1_name} + {self.x2_name} + ' \
-                  f'I({self.x1_name}**2) + I({self.x2_name}**2) + I({self.x3_name}**2)'
+                  f'{self.x1_name}_sq + {self.x2_name}_sq + {self.x3_name}_sq'
         self.model_simplified = smf.ols(formula, data=self.data).fit()
         print("\nModelo Simplificado:")
         print(self.model_simplified.summary())
@@ -117,9 +126,9 @@ class RSM_BoxBehnken:
         Ajusta un modelo Ridge Regression.
         """
         # Preparar datos
-        # Asegurarse de que los nombres de las columnas coincidan con los que se usaron en el modelo
+        # Usar los nombres consistentes de las columnas cuadráticas
         X = self.data[[self.x1_name, self.x2_name, self.x3_name, 
-                      f'I({self.x1_name} ** 2)', f'I({self.x2_name} ** 2)', f'I({self.x3_name} ** 2)']]
+                      f'{self.x1_name}_sq', f'{self.x2_name}_sq', f'{self.x3_name}_sq']]
         y = self.data[self.y_name]
         
         # Ajustar modelo Ridge
@@ -142,9 +151,9 @@ class RSM_BoxBehnken:
         Ajusta un modelo LASSO Regression.
         """
         # Preparar datos
-        # Asegurarse de que los nombres de las columnas coincidan con los que se usaron en el modelo
+        # Usar los nombres consistentes de las columnas cuadráticas
         X = self.data[[self.x1_name, self.x2_name, self.x3_name, 
-                      f'I({self.x1_name} ** 2)', f'I({self.x2_name} ** 2)', f'I({self.x3_name} ** 2)']]
+                      f'{self.x1_name}_sq', f'{self.x2_name}_sq', f'{self.x3_name}_sq']]
         y = self.data[self.y_name]
         
         # Ajustar modelo LASSO
@@ -167,9 +176,9 @@ class RSM_BoxBehnken:
         Ajusta un modelo ElasticNet Regression.
         """
         # Preparar datos
-        # Asegurarse de que los nombres de las columnas coincidan con los que se usaron en el modelo
+        # Usar los nombres consistentes de las columnas cuadráticas
         X = self.data[[self.x1_name, self.x2_name, self.x3_name, 
-                      f'I({self.x1_name} ** 2)', f'I({self.x2_name} ** 2)', f'I({self.x3_name} ** 2)']]
+                      f'{self.x1_name}_sq', f'{self.x2_name}_sq', f'{self.x3_name}_sq']]
         y = self.data[self.y_name]
         
         # Ajustar modelo ElasticNet
@@ -274,7 +283,10 @@ class RSM_BoxBehnken:
             return -self.model_simplified.predict(pd.DataFrame({
                 self.x1_name: [x[0]],
                 self.x2_name: [x[1]],
-                self.x3_name: [x[2]]
+                self.x3_name: [x[2]],
+                f'{self.x1_name}_sq': [x[0]**2],
+                f'{self.x2_name}_sq': [x[1]**2],
+                f'{self.x3_name}_sq': [x[2]**2]
             })).values[0]
         
         bounds = [(-1, 1), (-1, 1), (-1, 1)]
@@ -319,7 +331,10 @@ class RSM_BoxBehnken:
             return -self.model_simplified.predict(pd.DataFrame({
                 self.x1_name: [x[0]],
                 self.x2_name: [x[1]],
-                self.x3_name: [x[2]]
+                self.x3_name: [x[2]],
+                f'{self.x1_name}_sq': [x[0]**2],
+                f'{self.x2_name}_sq': [x[1]**2],
+                f'{self.x3_name}_sq': [x[2]**2]
             })).values[0]
         
         # Ejecutar optimización bayesiana
@@ -363,7 +378,10 @@ class RSM_BoxBehnken:
             return -self.model_simplified.predict(pd.DataFrame({
                 self.x1_name: [x[0]],
                 self.x2_name: [x[1]],
-                self.x3_name: [x[2]]
+                self.x3_name: [x[2]],
+                f'{self.x1_name}_sq': [x[0]**2],
+                f'{self.x2_name}_sq': [x[1]**2],
+                f'{self.x3_name}_sq': [x[2]**2]
             })).values[0]
         
         # Parámetros PSO
@@ -493,7 +511,7 @@ class RSM_BoxBehnken:
         
         # Variables predictoras
         X = self.data[[self.x1_name, self.x2_name, self.x3_name, 
-                      f'I({self.x1_name} ** 2)', f'I({self.x2_name} ** 2)', f'I({self.x3_name} ** 2)']]
+                      f'{self.x1_name}_sq', f'{self.x2_name}_sq', f'{self.x3_name}_sq']]
         
         # Calcular VIF para cada variable
         vif_data = pd.DataFrame()
@@ -629,7 +647,7 @@ class RSM_BoxBehnken:
         
         # Preparar datos
         X = self.data[[self.x1_name, self.x2_name, self.x3_name, 
-                      f'I({self.x1_name} ** 2)', f'I({self.x2_name} ** 2)', f'I({self.x3_name} ** 2)']]
+                      f'{self.x1_name}_sq', f'{self.x2_name}_sq', f'{self.x3_name}_sq']]
         y = self.data[self.y_name]
         
         # Calcular R² de validación cruzada
@@ -684,6 +702,10 @@ class RSM_BoxBehnken:
             varying_variables[1]: y_grid_coded.flatten(),
         })
         prediction_data[fixed_variable] = self.natural_to_coded(fixed_level, fixed_variable)
+        # Agregar términos cuadráticos
+        prediction_data[f'{varying_variables[0]}_sq'] = prediction_data[varying_variables[0]] ** 2
+        prediction_data[f'{varying_variables[1]}_sq'] = prediction_data[varying_variables[1]] ** 2
+        prediction_data[f'{fixed_variable}_sq'] = prediction_data[fixed_variable] ** 2
         
         # Calcular los valores predichos
         z_pred = self.model_simplified.predict(prediction_data).values.reshape(x_grid_coded.shape)
@@ -777,6 +799,10 @@ class RSM_BoxBehnken:
             varying_variables[1]: y_grid_coded.flatten(),
         })
         prediction_data[fixed_variable] = self.natural_to_coded(fixed_level, fixed_variable)
+        # Agregar términos cuadráticos
+        prediction_data[f'{varying_variables[0]}_sq'] = prediction_data[varying_variables[0]] ** 2
+        prediction_data[f'{varying_variables[1]}_sq'] = prediction_data[varying_variables[1]] ** 2
+        prediction_data[f'{fixed_variable}_sq'] = prediction_data[fixed_variable] ** 2
         
         # Calcular los valores predichos
         z_pred = self.model_simplified.predict(prediction_data).values.reshape(x_grid_coded.shape)
@@ -964,11 +990,11 @@ class RSM_BoxBehnken:
                     equation += f" + {coef:.3f}*{self.x2_name}"
                 elif term == f'{self.x3_name}':
                     equation += f" + {coef:.3f}*{self.x3_name}"
-                elif term == f'I({self.x1_name} ** 2)':
+                elif term == f'{self.x1_name}_sq':
                     equation += f" + {coef:.3f}*{self.x1_name}^2"
-                elif term == f'I({self.x2_name} ** 2)':
+                elif term == f'{self.x2_name}_sq':
                     equation += f" + {coef:.3f}*{self.x2_name}^2"
-                elif term == f'I({self.x3_name} ** 2)':
+                elif term == f'{self.x3_name}_sq':
                     equation += f" + {coef:.3f}*{self.x3_name}^2"
         
         return equation
@@ -1052,11 +1078,11 @@ class RSM_BoxBehnken:
         for index, row in anova_table.iterrows():
             if index != 'Residual':
                 factor_name = index
-                if factor_name == f'I({self.x1_name} ** 2)':
+                if factor_name == f'{self.x1_name}_sq':
                     factor_name = f'{self.x1_name}^2'
-                elif factor_name == f'I({self.x2_name} ** 2)':
+                elif factor_name == f'{self.x2_name}_sq':
                     factor_name = f'{self.x2_name}^2'
-                elif factor_name == f'I({self.x3_name} ** 2)':
+                elif factor_name == f'{self.x3_name}_sq':
                     factor_name = f'{self.x3_name}^2'
                 ss_factor = row['sum_sq']
                 df_factor = row['df']
@@ -1100,7 +1126,7 @@ class RSM_BoxBehnken:
         # --- ANOVA detallada ---
         # 1. Ajustar un modelo solo con los términos de primer orden y cuadráticos
         formula_reduced = f'{self.y_name} ~ {self.x1_name} + {self.x2_name} + {self.x3_name} + ' \
-                         f'I({self.x1_name}**2) + I({self.x2_name}**2) + I({self.x3_name}**2)'
+                         f'{self.x1_name}_sq + {self.x2_name}_sq + {self.x3_name}_sq'
         model_reduced = smf.ols(formula_reduced, data=self.data).fit()
         
         # 2. ANOVA del modelo reducido
@@ -1158,9 +1184,9 @@ class RSM_BoxBehnken:
         })
         
         # Calcular la suma de cuadrados y estadísticos F para la curvatura
-        ss_curvature = anova_reduced['sum_sq'][f'I({self.x1_name} ** 2)'] + \
-                      anova_reduced['sum_sq'][f'I({self.x2_name} ** 2)'] + \
-                      anova_reduced['sum_sq'][f'I({self.x3_name} ** 2)']
+        ss_curvature = anova_reduced['sum_sq'][f'{self.x1_name}_sq'] + \
+                      anova_reduced['sum_sq'][f'{self.x2_name}_sq'] + \
+                      anova_reduced['sum_sq'][f'{self.x3_name}_sq']
         df_curvature = 3
         ms_curvature = ss_curvature / df_curvature
         f_curvature = ms_curvature / ms_residual