We had yet another very informative and well presented lecture on
outlier detection last week by Bahjat. Sorry for the on again off again
nature of that meeting.

I have tried to summarize below some of the suggestions I've received
this week. The first two articles with abstracts have pdf's so Lisa I
can send any to you that are not already posted.

Doug Kelly and I are planning to talk about next semester's meetings.
Doug will take an active role in convening the group as I have been
doing this time and I know, Doug, you expressed interest in computer
anomaly detection - see Kevin's suggestions below. If that is a
direction we want to go, I have names of some folks at State who have a
group focused on this issue and I think early on, we were going to check
with Duke and UNC.

David had mentioned possibly getting up with Prof. Priebe who has that
SCAN statistics on Enron Graphs listing in our Sept. 29 web page section.

Some of us are working together on projects or talking about them. It
helps SAMSI if we reference them in any publications that result.

These are just some thoughts to start discussion.

Dave

******************************************************************
From Francisco

Spatial Forecast Methods for Terrorist Events in Urban Environments

Donald Brown, Jason Dalton, and Heidi Hoyle

Abstract. Terrorist events such as suicide bombings are rare yet
extremely destructive events. Responses to such events are even
rarer, because they require forecasting methods for effective
prevention and early detection. While many forecasting methods
are available, few are designed for conflict scenarios. This
paper builds on previous work in forecasting criminal behavior
using spatialchoice models. Specifically we describe the fusion
of two techniques for modeling the spatial choice of suicide bombers
into a unified forecast that combines spatial likelihood modeling
of environmental characteristics with logistic regression
modeling of demographic features. ...

Criminal incident prediction using a point-pattern-based density
model

Hua Liu , Donald E. Brown
CSG Systems , Inc .,
One Main Street ,
Cambridge ,MA 02142,USA

Abstract
Law enforcement agencies need crime forecasts to support their tactical
operations; namely, predicted crime locations for next week based
on data from the previous week. Current practice simply assumes that
spatial clusters of crimes or ‘‘hot spots’’ observed in the previous
week will persist to the next week. This paper introduces a multivariate
prediction model for hot spots that relates the features in an area to
the predicted occurrence of crimes through the preference structure of
criminals.We use a point-pattern-based transition density model for
space–time event prediction that relies on criminal preference discovery
as observed in the features chosen for past crimes. The resultant
model outperforms the current practices, as demonstrated statistically
by an application to breaking and entering incidents in Richmond,VA.

Bahjat and Jim sent articles we mentioned a couple of weeks ago.

From Kevin

1. Wegman and Marchette (2003),
On Some Techniques for Streaming Data:
A Case Study of Internet Packet Headers

2. Wegman and Marchette (2003),
Statistical Analysis of Network Data for Cybersecurity

3. Marchette (1999),
A Statistical Method for Profiling Network Traffic

Gallit's BLOG
(http://www.smith.umd.edu/faculty/gshmueli/blog/)

Upcoming: Thanksgiving, Dave (unit roots Dec. 1),
Curt(P-scan article by Naus and Wallenstein Dec. 8)
This takes us to Dec. 8. Do we want to meet
on Dec. 15? I'd assume not on Dec. 22 through
the first of the year (?)

Lingshong Zhang has volunteered to present.

Myron has volunteered to say something about
threshold autoregression.