Welcome to the Anomaly Detection Model Build Wizard

The welcome page of the wizard contains a brief description of anomaly detection models, which use the Support Vector Machine algorithm to detect values that are atypical.

One-class SVM models are useful in situations where you wish to detect atypical or anomalous cases, that is, cases that don't fit with the majority of cases. For example, you might want to predict people likely to purchase a product when you have detailed information about people who did purchase the product and have little or no information about people who did not purchase the product.

You can also build anomaly dectection models in case where you do have both examples and counterexamples for the rare situation. If you have counterexamples, you should consider building a classification model using the SVM algorithm; in some case, the classification model may give better results.

For more information about anomaly detection, see Anomaly Detection.

If you do not want to see this page the next time that you launch the wizard, check the Skip this Page Next Time box.

Click Next to continue.