This topic describes how to evaluate a model. A binary classification model is used as an example in this topic.

Prerequisites

A model is generated. For more information, see Generate a model.

Procedure

  1. Log on to the Machine Learning Platform for AI and navigate to the pipeline page.
    For more information, see Prepare and preprocess data.
  2. Create a prediction node.
    Prediction
    1. In the list of components, search for the Prediction component, drag and drop the component to the canvas, and then specify the generated node as the child node of the Split-1 node and the Logistic Regression for Binary Classification-1 node.
    2. Click the Prediction-1 node on the canvas. In the right-side panel that appears, click Select Fields in Feature Columns and Reserved Columns. In the Select Fields dialog box, click Edit in the Selected section and set the following parameters.
      • Feature Columns: In the code editor, enter the following fields except status: age,sex,cp,trestbps,chol,fbs,restecg,thalach,exang,oldpeak,slop,ca,thal,style.
      • Reserved Columns: Enter status in the code editor.
  3. Create a node named Binary Classification Evaluation.
    Prediction evaluation
    1. In the list of components, search for the Binary Classification Evaluation component, drag and drop the component to the canvas, and then specify the generated node as the child node of the Prediction-1 node.
    2. Click the Binary Classification Evaluation-1 node on the canvas. On the right-side Fields Setting tab, set Original Label Column to status.
  4. In the upper-left corner of the canvas, click Run.
  5. View the evaluation report of the model.
    1. After the node stops running, right-click the Binary Classification Evaluation component. In the shortcut menu that appears, click View Analysis.
    2. Click the Evaluation Charts tab and view the receiver operating characteristic (ROC) curve of the binary classification model with different parameters. Model evaluation