Pawan Kumar Pandey

Title: Pawan Kumar Pandey and Provider Activity Anomaly Detection.

Introduction

Pawan Kumar Pandey is an inventor based in Karnataka, India, and works at Optum, Inc. His patent work addresses the use of unsupervised learning to identify provider addresses that may have become inactive.

Latest Patents

Pandey is associated with 1 patent, titled “Provider Activity Anomaly Detection.” The invention describes techniques for automatically detecting inactive provider addresses by analyzing activity trends. When an address shows increasing activity, the system determines an active metric that estimates the likelihood that the address is active. When activity is decreasing, it determines an inactive metric to estimate the likelihood that the address is inactive. The system then uses these metrics to classify the address as active or inactive.

In some embodiments, the metrics are calculated as weighted sums of z-scores derived from Gaussian distributions based on benchmarks or provider features. The patent also describes using reinforcement learning from human feedback (RLHF) techniques to determine the weights.

Career Highlights

Pandey’s listed invention applies machine-learning methods to provider data, with the goal of identifying changes in address activity. The approach combines activity trends and statistical metrics to support automated classification.

Collaborations

Pandey’s listed coworkers include Ramesh Rangasamy Ganesan and Kevin Obrien.

Conclusion

Pawan Kumar Pandey’s patent on provider activity anomaly detection describes a data-driven approach to assessing whether provider addresses remain active, combining trend analysis, statistical measures, and weighted metrics.