Linear model feature importance

Limitations

The supported computing engine is MaxCompute.

Component configuration

You can configure the parameters for the Linear Model Feature Importance component in one of the following ways.

Method 1: Visual interface

Method 2: PAI command

PAI -name regression_feature_importance -project algo_public
    -DmodelName=xlab_m_logisticregressi_20317_v0
    -DoutputTableName=pai_temp_2252_20321_1
    -DlabelColName=y
    -DfeatureColNames=pdays,previous,emp_var_rate,cons_price_idx,cons_conf_idx,euribor3m,nr_employed,age,campaign
    -DenableSparse=false -DinputTableName=pai_dense_10_9;

Example

1. Create a table named bank_data and import data into it. For more information, see Create a table and Import data.

2. Run the following SQL statement to generate training data.

   create table if not exists pai_dense_10_9 as
   select
       age,campaign,pdays, previous, emp_var_rate, cons_price_idx, cons_conf_idx, euribor3m, nr_employed, fixed_deposit
   from  bank_data limit 10;

3. Build and run a pipeline. For more information, see Algorithm modeling. To connect the components, connect the output of Read Table-1 to the inputs of both Logistic Regression for Multiclass Classification and Linear Model Feature Importance. Then, connect the output of Logistic Regression for Multiclass Classification to the input of Linear Model Feature Importance.

   1. Drag the Read Table, Logistic Regression for Multiclass Classification, and Linear Model Feature Importance components from the component list in Machine Learning Designer to the canvas.

   2. Connect the components to build the pipeline.

   3. Configure the parameters for each component.

      • On the canvas, click the Read Table-1 component. In the right-side pane, on the Select Table tab, set Table Name to bank_data.

      • On the canvas, click the Logistic Regression for Multiclass Classification-1 component. In the right-side pane, on the Fields Setting tab, set Feature Columns to age, campaign, pdays, previous, emp_var_rate, cons_price_idx, cons_conf_idx, euribor3m, and nr_employed. Set Target Column to fixed_deposit. Use the default values for the other parameters.

      • On the canvas, click the Linear Model Feature Importance-1 component. In the right-side pane, on the Fields Setting tab, set Target Column to fixed_deposit. Use the default values for the other parameters.

   4. After you configure the parameters, click the Run button .

4. After the pipeline runs successfully, right-click the Linear Model Feature Importance-1 component and choose View Data > Model Importance Table. The Linear Model Feature Importance component outputs a table that shows the colname, weight, and importance for each feature. Values are presented in scientific notation. For example, the weight of the age feature is 9.61816270808075E-5.

   The metric formulas are as follows.

   Column name	Formula
   weight	abs(w_)
   importance	abs(w_j) * STD(f_i) Note abs(w_j) is the absolute value of the feature coefficient, and STD(f_i) is the standard deviation of the training data.

5. Right-click the Linear Model Feature Importance-1 component and select View Analytics Report.