Learning and Predicting Patent Technology Reuse Trajectories from Emergence-Time Signals

Forecasting how a newly emerged patent technology will be reused is central to technology intelligence, but reuse-pattern labels do not exist in advance: they must be constructed from the trajectories themselves, and how they are constructed determines what a forecast means. We study $201{,}710$ nov