In areas of induced seismicity, earthquakes can be triggered by stress changes from fluid injection and from static deformation caused by fault slip. Here we present a method to distinguish between injection-driven and earthquake-driven triggering of induced seismicity by combining a calibrated, fully-coupled, poroelastic stress model of wastewater injection with a random forest machine learning algorithm trained on both earthquake catalog and modeled stress features. We investigate the classic Paradox Valley, Colorado induced seismicity dataset as an ideal test case: a single, high-pressure injector that has induced >7000 earthquakes between 1991 and 2012. We find that injection-driven earthquakes are approximately 22±-5% of the total catalog and have distinct spatiotemporal clustering with a larger b-value, closer proximity to the well and earlier occurrence in the injection history. Our model may be applicable to other regions to help determine site’s susceptibility to triggered earthquakes due to fluid injection.