Limit to 6 features

app.py

@@ -6,7 +6,7 @@ from sklearn.linear_model import LinearRegression
 from sklearn.ensemble import RandomForestRegressor
 from sklearn.datasets import fetch_california_housing
 
-def trainme(MedInc, HouseAge, AveRooms, AveBedrms, Population, AveOccup, Latitude, Longitude, Price):
+def trainme(MedInc, AveOccup, HouseAge, Latitude, Longitude, Price):
 
   #Call base dataset
   housing = fetch_california_housing()
@@ -14,36 +14,29 @@ def trainme(MedInc, HouseAge, AveRooms, AveBedrms, Population, AveOccup, Latitud
   #As a dataframe, with tragets
   housing_df = pd.DataFrame(data=housing.data, columns=housing.feature_names)
   housing_df["Price"] = housing.target
+  housing_df = housing_df.drop(['AveBedrms', 'Population', 'AveRooms'], axis=1)
 
   #Put the new line at index 0
-  housing_df.iloc[0] = [MedInc, HouseAge, AveRooms, AveBedrms, Population, AveOccup, Latitude, Longitude, Price]
+  housing_df.iloc[0] = [MedInc, AveOccup, HouseAge, Latitude, Longitude, Price]
 
   #Split
   y = pd.DataFrame(data=housing_df['Price']).to_numpy()
-  X = housing_df.drop('Price', axis=1).to_numpy()
+  X = housing_df[['MedInc','AveOccup','HouseAge', 'Latitude', 'Longitude']].to_numpy()
 
   #create a machine learning model and train it
   regressor = RandomForestRegressor(n_estimators = 10, random_state = 0)
-  regressor.fit(X,y)
+  regressor.fit(X,np.ravel(y))
 
   #Create a redable/clean feature list
-  clean_features = ['Median income',
- 'Median house age',
- 'Average number of rooms per household',
- 'Average number of bedrooms per household',
- 'Population of the block',
- 'Average number of household members',
- 'Latitude of the block',
- 'Longitude of the block']
+  clean_features = ['Median income','Average number of household members','Median house age','Latitude', 'Longitude']
 
   #Create a lime object, regression mode
   from lime import lime_tabular
   explainer = lime_tabular.LimeTabularExplainer(X, mode="regression", feature_names=clean_features)
 
   #Create the expl
-  explanation = explainer.explain_instance(X[0], regressor.predict, num_features=len(housing.feature_names))
+  explanation = explainer.explain_instance(X[0], regressor.predict, num_features=5)
   listing = explanation.as_list()
-  #explanation.show_in_notebook(show_table=True)
 
   #Get pred and actual scores
   Pred_value = regressor.predict(X[0].reshape(1,-1))*100000
@@ -159,17 +152,14 @@ def trainme(MedInc, HouseAge, AveRooms, AveBedrms, Population, AveOccup, Latitud
 
   return final_list
 
-#Define gradio UI
+#Define gradio UI MedInc, AveOccup, HouseAge, Latitude, Longitude, Price):
 
 input = [gr.inputs.Slider(0.5, 15, default=4.406300, label='Median income'),
-         gr.inputs.Slider(1, 52, default=13, label='Median house age'),
-         gr.inputs.Slider(1, 141, default=28.757282, label='Average number of rooms per household'), 
-         gr.inputs.Slider(1, 34, default=5.300971, label='Average number of bedrooms per household'),
-         gr.inputs.Slider(3, 36000, default=252, label='Population of the block'),
-         gr.inputs.Slider(1, 1200, default=2.446602, label='Average number of household members'),
-         gr.inputs.Slider(32, 42, default=38.260000, label='Latitude of the block'),
-         gr.inputs.Slider(-124, -111, default=-120.330000, label='Longitude of the block'),
-         gr.inputs.Slider(0.14, 5, default=1.558000, label='Price'),
+        gr.inputs.Slider(1, 30, default=2.446602, label='Average number of household members'),
+        gr.inputs.Slider(1, 52, default=13, label='Median house age'),
+        gr.inputs.Slider(32, 42, default=38.260000, label='Latitude of the block'),
+        gr.inputs.Slider(-124, -111, default=-120.330000, label='Longitude of the block'),
+        gr.inputs.Slider(0.14, 5, default=1.558000, label='Price'),
          ]
 output = [gr.outputs.Textbox(label='Prediction'),
           gr.outputs.Textbox(label='Why?'),
@@ -182,4 +172,4 @@ iface = gr.Interface(fn=trainme,
                      outputs=output,
                      title = '🔮 Explain me like I\'m 5',
                      description="Get explanations from a model prediction")
-iface.launch(debug=True)
+iface.launch()