A Poisson process has rate of N per unit length per unit time. The events occur uniformly on the x-axis but I always have measurement error of greater than 1/N. I was wondering is it possible that I can use my past measurements to reduce the error on my current measurement ? i.e. If I have collection of confidence intervals [x_min, x_max] as my data is it possible to use this information and reduce the measurement error in the current measurement.