The Analysis of Cyclic Point Processes with Walsh Functions.
Abstract
Transformation by Walsh functions is a technique useful for uncovering the interactions among the daily, weekly, and seasonal cycles often present in societal data. The structural and sampling properties needed for interpretation are simple, the former because Walsh functions are easy to visualize, the latter because the normal approximation applies whenever the total number of events is large. This paper demonstrates these properties, extends this technique to multivariate point processes, and discusses display of the counting function, a technique which complements the Walsh transform. These techniques are illustrated by applying them to a series of robberies. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1975
- Accession Number
- ADA018893
Entities
People
- Walter S. Liggett Jr.