Outlier Detector™ from Elite Analytics is an anomaly detection reporting tool used by business and data analysts to perform sophisticated outlier analysis without programming. Outlier Detector™ combines the process of building an anomaly detection model with a unique and easy-to-use desktop interface to allow analysts to explore their model results and generate visualizations that are easily understood. Features include:
Data Import and Browser
Easily connect to source files or databases to import your dataset. Handy data viewer provides easy access to view, sort and query your source data as you build your model metrics.
The Model Builder provides a drag-and-drop user interface to quickly define data set groupings and the metrics used to score each group. Currently supported metrics include simple, ratio and statistic. Metric Controls let you define custom categorical or numeric values to adjust metric computations for explainable variations such age, income, or geography. Define a Composite Score to build an overall outlier metric to simplify exploratory analysis. If you are experienced in using “R”, the Model Builder provides a Custom Metric for you to define the most complex metric imaginable.
The Results Visualizer provides an intuitive, graphical interface for exploring your model results. Start with a list of groups and their corresponding metrics and metric scores to identify outliers. A “stoplight” indicator categorizes outliers into green (normal), yellow (outlier), and red (extreme outlier). Box plot, histogram and scatter-plot graphs can be generated with a simple mouse-click and are linked to your results list to highlight specific outliers in the graphs. Click on any group and drill-down into the transaction details that were used to compute the metrics.
The algorithms and analysis techniques used in Outlier Detector™ were developed by Elite Analytics over many years of successfully applying outlier analysis in different industries. The product packages our expertise into the modeling engine so business analysts can concentrate on research and investigation rather than tedious database and statistical programming.
If your business situation has underlying transactional data organized by any sort of natural grouping that you want to understand in more detail, Outlier Detector™ can help you find the anomalies in your data.
Example industry applications include:
- Tax Return
- Anti-money laundering
- Stock trading
- Insider Hacking
- Auto Warranty