A method is provided for characterizing sparse data as either signal or noise. In one embodiment, a two-dimensional area which contains a distribution of data points of the sparse data from a selected time period is divided into equal size cells wherein each cell has an expectation of containing at least one data sample. Based on the expected proportion of a plurality cells which will be nonempty in a random distribution, a Poisson distribution or a Binomial distribution is utilized to determine a data sample mean. A probability of a false signal is determined from the number of cells utilized. A probability is also computed from the sample mean and compared to the probability of a false signal to determine whether to characterize the sparse data as signal or noise.

 
Web www.patentalert.com

> Optical scanner

~ 00376