PREDICTIVE DATA MINING
New relationships. Unique visualization.
Our data scientists use a suite of applications – regression, decision trees, neural networks – along with our proprietary machine learning software to mine data for new, non-intuitive relationships.
This software utilizes a data mining/modeling technique called genetic programming to test every possible variable combination in search of new predictive incite.
This testing process automatically builds 1,000 models per second, saves the best model, integrates it into the next generation of 1,000 models and repeats the testing process. At the completion of 100,000 models, the variables that “survive” this testing process make up the final model algorithm. The code from that model algorithm is then used to score (rank) client records. We call our data mining/modeling product Model Incite.
We offer our Model Incite product on a project or subscription based on your specific needs.