When weighting the past evens, e.g., with an exponential decay, I think we want to 'start counting' from the time of the previous event rather than the current event (which I think is being done now). This ensure that the constant hazard assumption is not violated.
When weighting the past evens, e.g., with an exponential decay, I think we want to 'start counting' from the time of the previous event rather than the current event (which I think is being done now). This ensure that the constant hazard assumption is not violated.