|
Activity Number:
|
475
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
IMS
|
| Abstract - #305062 |
|
Title:
|
Testing for Independence Between a Time Series and a Point Process
|
|
Author(s):
|
Victor Solo*+
|
|
Companies:
|
University of New South Wales
|
|
Address:
|
School of Electrical Engineering, Sydney, 2052, Australia
|
|
Keywords:
|
point process ; stochastic intensity
|
|
Abstract:
|
In an increasing number of applications data arises which consists of both point process and analog (i.e. continuous valued) process. This occurs in neural coding where electrodes inserted in the brains of animals record both spike trains and so-called local field potentials; it also occurs in electricity pricing where peak demand can generate sudden price spikes so that the overall price is best modeled as a sum of a point process and an analog price process. However there is almost no literature devoted to joint analysis of point processes and analog processes. Here we review the scant literature, show how to construct a joint likelihood and use this to develop a test of independence.
|
