PyODPS feature scaling

Procedure

1. Download the test dataset and import it into MaxCompute.

   1. Download and decompress the Iris dataset. Rename the iris.data file to iris.csv.

   2. Create a table named pyodps_iris and upload the iris.csv dataset. For more information, see Create a table and upload data.

      Use the following statement to create the table.

      CREATE TABLE if not exists pyodps_iris
      (
      sepallength  DOUBLE comment 'Sepal length (cm)',
      sepalwidth   DOUBLE comment 'Sepal width (cm)',
      petallength  DOUBLE comment 'Petal length (cm)',
      petalwidth   DOUBLE comment 'Petal width (cm)',
      name         STRING comment 'Species'
      );

2. Log in to the DataWorks console and select a region in the upper-left corner.

3. In the left-side navigation pane, click Workspaces.

4. On the Data Development page, right-click the created workflow and choose Create Node > MaxCompute > PyODPS 2.

5. In the Create Node dialog box, enter a node name and click OK.

6. In the PyODPS 2 node, enter the feature scaling code.

   Example code:

   # Load the table into a DataFrame.
   df = DataFrame(o.get_table('pyodps_iris'))
   
   # Rename the 'petallength' column to 'fid' to align with the example's feature column.
   df = df.rename(columns={'petallength': 'fid'})
   
   # Use the min_max_scale method for normalization to the default range of [0, 1].
   print df.min_max_scale(columns=['fid']).head()
   
   # Use the feature_range parameter to specify a custom output range, such as [-1, 1].
   print df.min_max_scale(columns=['fid'],feature_range=(-1,1)).head()
   
   # Use the preserve parameter to retain the original column.
   # The scaled data is added as a new column with the `_scaled` suffix.
   print df.min_max_scale(columns=['fid'],preserve=True).head()
   
   # Use the group parameter to perform scaling separately within each group.
   print df.min_max_scale(columns=['fid'],group=['name']).head()
   
   # Use the std_scale method for standardization (zero mean and unit variance).
   print df.std_scale(columns=['fid']).head()

7. Click Run.

8. On the Run Log tab, view the results.