Classification Learner is an app in the Statistics and Machine Learning Toolbox™ that trains models to classify data. Classification Learner uses automated methods to create and test different types of classification models using labeled datasets. For predictive maintenance, the goal of using Classification Learner is to select and train a model that uses the exported feature set to discriminate between data from healthy and from faulty systems. You can incorporate this model into an algorithm for fault detection and prediction.
Use Export features to Classification Learner when:
You want to obtain more insight on the relative effectiveness of your features.
You are developing a predictive maintenance algorithm and want to select and train the best model for the algorithm to use.
When you export features, the app brings up a selectable list of features to export. The specific list depends on where you execute the export.
If you export from the Feature Designer tab, the list is in alphabetical order, with all features preselected. This approach allows you to export all features at once from the main tab. You can also tailor the selections at this level if you know which features you want.
If you export from the Feature Ranking tab, the list is in ranked order, based on the ranking method in Features sorted by. In this ranked list, the top five features are preselected. This approach allows you to export only your highest-ranked features. You can also tailor the selections at if you want to export more than the top five features.
For more information, see Classification Learner.