This is a proposal to enhance the phase_dispersion_search to allow for different weighting of observations (events in the light curve). Such an enhancement would open up using the phase_dispersion_search for light curves with upper limit detections (as needed in my current research), and could otherwise improve the sensitivity of PDM by enabling down weighting of uncertain (possibly spurious) observations.

To this end, a new variances parameter can be introduced for the phase_dispersion_search function that is used for the observation weights: event_weights = 1/variances.