STATISTICA Blogs

Recently I sat down with Jonathan Polon and he took a moment to discuss his use of STATISTICA. Jonathan Polon is one of the founders of Claim Analytics. They provide in-depth expertise in advanced predictive modeling to major insurance companies to assist with claim scoring, pricing, reserving and fraud detection.

1. What is your background (career, education, etc.)?
   Bachelor or Sciences in Actuarial Mathematics, University of Michigan, 1994. Fellow of the Society of Actuaries, 1999.
    
2. How do you use STATISTICA? Which modules?
   We use STATISTICA to build predictive models for Life, Health and P&C insurance companies for a variety of applications such as claims management, fraud detection, ratemaking and underwriting. We primarily work with the Data Mining and Statistics Modules and have also tried Text Mining.
    
3. What are some interesting consulting projects that you’ve used STATISTICA for?
   The most interesting project was using unsupervised learning for medical provider fraud detection. It was very challenging to try and find something when we didn’t know what we were looking for. Another interesting project was applying text mining for predicting the severity of workers compensation claims and to see that a few key phrases in the adjuster notes were just as predictive as several complex measurements derived from the structured claims data.
    
4. What types of customers do you consult with (in other words what is your specialty)?
   My specialty is insurance and, even though I am an actuary, my focus is building and applying predictive models to support claims management processes.
    
5. What advantages do you see in STATISTICA compared to other general tools?
   We see a few key advantages with STATISTICA as compared to other general tools:
   • Comprehensiveness of Predictive Modeling Algorithms: STATISTICA offers a very complete catalogue of predictive modeling algorithms from all families (eg, regression, trees, neural nets, machine learning, etc). Also, new algorithms tend to be made available in STATISTICA earlier than in competing products (eg, ensemble methods and SVM).
   • Working with StatSoft: StatSoft is very responsive and supportive. Whether we have technical questions, licensing questions or want an opportunity to trial a new module, everyone at StatSoft is very easy to work with.
   • Scalable: When we were a startup, STATISTICA was well-suited to our needs. The single user license was very affordable but still provided the complete set of functionality that we required and the support we received was superb in spite of our being a small account. As we continue to grow STATISTICA continues to have solutions that fit our company.
      
6. Would you recommend STATISTICA?
   Yes!