Machine Learning Example in ARPIA
Creating a Machine Learning Model with AutoML
In this example, you will learn how to quickly build a Machine Learning model with ARPIA's AutoML.
Step 1 From your Home page in the Orchestrator, select the Data Workshop.
Step 2 Choose the project where you want to create the Machine Learning model.
Step 3 Click on the + button.
Step 4 A new window will open. Select Machine Learning for the Type field, DK AutoML Engine for Development Environment. Then, enter a Name for your object and an optional description.
Step 5 Select the Engine Settings. Choose the preconfigured models you want to use, the percentage of data to use for testing the model, whether to randomize the dataset, in AutoML Engine Status select Metrics, and choose if you want to retrain on each Batch.
Classification and Regression Models
ARPIA's AutoML includes classification and regression models. To learn the difference, read the blog ¿Clasificación o Regresión?
Step 6 In Engine Dataset Source, choose the Table you will use to train your model. In Features Columns, select the columns you will use to train the model, and choose the column to predict in Target Column.
Step 7 Click on SAVE SETTINGS and then on RUN.
Step 8 When you click RUN, a new window will open. Click on RUN again to execute the model. In the status, you can observe if the model is training correctly, and in Results, the execution.
Step 9 After training the models, the Metrics section will open immediately, where you can observe the evaluation metrics for each model. Click on the model with the best accuracy.
Accuracy Metric
The accuracy metric helps us understand how well the model performed. The closer to 1, the better the model.
Step 10 In the ENGINE SETTINGS section and in AutoML Engine Status, select Deployed.
Step 11 Click on SAVE SETTINGS and then on RUN.
Step 12 You will again see the Metrics section, where you can observe the behavior of your model ready to use.
Step 13 In the PREDICT NEW DATA section, choose the table to which you want to apply the model. Click on SAVE SETTINGS and run it. This will store the predictions in the selected table.
Columns "predicted, predicted_dt" and "predicted_accuracy"
It is not necessary to add these columns manually. If the table does not have these columns, ARPIA will create them automatically.
This way, you have created your first AutoML model in ARPIA.
Updated about 1 month ago