Some of the Projects:
- Anomaly Detection algorithms to identify suspicious vision care providers for the Special Investigative Units (SIU) and Data-Warehousing groups – deriving fraud indicators, controlling the fraud indicators for known or explainable causes, developing outlier metrics at the provider level, and combining the multiple outlier metrics to derive a provider fraud score
- Machine Learning model to predict the probability that a member would purchase vision care services outside the insurance network. The model uses historical insurance claim patterns and member demographics as input. The model is automatically triggered to refresh the prediction for a given member when there is a change to underlying data
- Machine Learning model to predict “member churn,” that is, the probability that an existing member will not renew their vision care benefit
- Machine Learning cluster models to respond to business changes, especially those affected by Covid-19. The Strategy Office and Marketing Department use this solution for membership segmentation to reposition or create new programs, products, and meet the interests and needs of their current and potential members. This effort will lay the foundation for making this process more effective and repeatable internally.