MINNEAPOLIS – Fair Isaac & Co., the maker of the ubiquitous FICO score, announced yesterday it has been awarded eight new patents for inventions in fraud detection, search technology, rules management and bioanalytics.
The new patents include two technologies used in the new Fair Isaac Falcon Fraud Manager version 6, which produces a 44% improvement in fraud detection model performance, increased fraud detection and fewer false-positives.
The first, "Cascaded Profiles for Multiple Interacting Entities," relates to the generation of enhanced profiles used in fraud detection which powerfully reflect interactions and relationship of key entities. The second, "Self-Learning Real-Time Prioritization of Fraud Control Actions," uses adaptive models to learn new fraud patterns in real-time and by constantly re-weighting predictive fraud variables related to activity based on real-time case feedback.
The new portfolio also includes an additional patent in transactional profile analytics, entitled "Enhancing Delinquent Debt Collection Using Statistical Models of Debt Historical Information and Account Events." This patent describes an automated system using predictive modeling based on real-time profiles (common to Falcon fraud models) including the use of unstructured data to optimize the use of collection resources on a portfolio of delinquent debt accounts.
The company also received its first Japanese patent for "Context Vector Generation and Retrieval," adding foreign coverage to four Fair Isaac U.S. patents related to search tools. This invention is an improved system and method for generating and retrieving "context vectors," which are widely used search tools, including for Web-based searching.
Fair Isaac now holds 80 patents, nearly double the portfolio of two years ago. The company filed for 10 more patents in the past four months, and currently has 160 patent applications pending in the U.S. and other countries.
