FAB High Q‎ > ‎

About GT Data Mining

GT is a data mining analytics solution based on a strong CLUSTERING algorithm. It has a unique ability to identify patterns of behavior out of complex multi-variant data, including hidden and rare events.

GT offers to the analyst new and fresh view of the data functionality in a hierarchical format, indicating what are the characteristics of each pattern and if it is exclusive. This simple presentation of the groups enables focusing on each patterns and locating its key factors. Rolling up the groups uncovers data behavior at the root cause of events, the conditions that lead to the various types of events.
Drilling down the GT patterns reveals the variables which dominate each of the groups. We use the fact that some of the effective variables appear at an earlier stage in order to generate alerts and predictions.
 
GT output contains:
  • group hierarchy
  • group characteristics  
  • rules of behavior
  • prediction formulas
  • indicators
  • charts and tables
  • update routine
  • interpretation of causes and root causes.
 

GT robust algorithms can overcome input imperfections and "noise" thanks to the group structure which balances out the random occurrences.

The solution has innate tests that cross check the results before submitting them to the client.

GT's new discoveries and hypotheses can be later validated with statistics.

This new solution is highly cost-effective, as results of using existing data, and demanding minimal organization overheads while providing maximum added value.