6.1 Introduction

Besides finding out the anomalous instances in a dataset, you can also use your anomaly detector to score new data that the model has not yet seen. Predictions for anomalies are referred to as anomaly scores in BigML, since they aim to quantify the level of anomalousness for new data instances. Anomaly scoring is possible either for single instances, i.e., one by one, or for multiple instances simultaneously, i.e., in batch. Each score comes with a field importance measure to indicate the relative contribution of each field in the anomaly score.

The predictions tab in the main menu of the BigML Dashboard is where all of your saved predictions are listed (Figure 6.1 ). In the scores list view, you can see the icon for the Anomaly Detector used for each score, the Name of the score, the Anomaly Score, and the Age (time since the score was created). You can also search your scores by name clicking in the search menu option on the top right menu.

By default, when you first create an account at BigML, or every time that you start a new Project, your list view for predictions will be empty. (See Figure 6.2 .)

Anomaly scores are saved under the Anomaly Detection otpion in the menu (see Figure 6.3 .)

From this view, you can select to view the list of your single anomaly scores or your batch anomaly scores by clicking on the corresponding icons (see Figure 6.4 and Figure 6.5 .)