Regression is essentially a statistical approach used to find the relationship between variables. In machine learning, this is used to predict the outcome of an event based on the relationship between variables obtained from the dataset.
As we've seen with prior options for training a model, the product documentation gives us a very good example exercise we can use to illustrate the regression approach to machine learning: training a model to predict the amount of money a customer is likely to spend on a trip to an outdoor equipment store.
Again, we'll go over the appropriate steps required for this exercise. For this exercise, we will choose the following:
- PURCHASE_AMOUNT (which is the average amount of money the customer has spent on each visit to the store) as our label column
- GENDER, AGE, MARITAL_STATUS, and PROFESSION as our feature columns
Next, as in the preceding section's exercise, we will again click on Manual so that we will be able to choose the specific algorithms the model uses (instead of letting the model builder chose for us), then perform the following steps:
- For the Select a technique option (shown in the following screenshot), select Regression.
- Add the estimator named Gradient Boosted Tree Regressor. The following screenshot indicates the parameters we've chosen for this exercise model build:
- Once the preceding details are set, you can click on Next to begin training the model with the sample data, using the selected technique and estimator. After training completes, you can click on Save. Of course, after the model is saved, the model details page opens automatically:
- In the usual fashion, to verify the algorithm the model builder used, you can again go to the Summary table in the Overview information on the model details page (shown in the following screenshot) and click on View in the Model builder details row:
