Predicting the Future:
AI Approaches to Time-series Problems


Many dream of being able to predict the future. In finance, accurate predictions can direct portfolio management decisions. In marketing, predicting future demand for products and services can direct capital allocation.

When crystal balls are not available, one may rely on analysis of historical data to discover predictive patterns. Temporal patterns are of particular interest because of the large number of high-profile applications that include historical time series. The goal of this workshop is to bring together AI researchers who study time-series problems, along with practitioners and researchers from related fields, in order to establish common ground.


We are interested in original research results and application solutions involving the automated analysis of time-series data.

Authors are asked to address the following questions, where applicable:

* How have you formulated the time-series analysis problem?
  Do you build complete classification or regression
  models?  discover temporal patterns?

* Are you focusing on the creation of a new algorithm?  On the
  creation of temporally oriented features?

* Have you built on related work from AI or from other communities?

* Is your method designed for a particular application?  Do your results

* Have you evaluated your work?

Workshop Format

The workshop will be a one-day focused working session with several paper sessions, each followed by ample time for discussion. It will also include invited talks designed to portray time-series analysis from a wide range of perspectives (statistical, state space, etc.).


Attendance will be by invitation. 25-50 people will be invited based on experience, as indicated by paper submissions or research summaries.

Submission Requirements

Authors should submit papers of no more than 8 pages (12-pt font) to the workshop chair. Each submission should include authors' e-mail and surface mail addresses, a short abstract, and a list of keywords. E-mailed submissions of Postscript files are preferred; if a Postscript file cannot be produced, submit four (4) printed copies.

Those interested in attending without submitting a paper should submit a one-page summary of related work.


 Submission deadline:               March 11, 1998
 Notification date:                 April 1, 1998
 Camera-ready copies due:           April 22, 1998

Organizing Committee

     Andrea Danyluk (workshop chair)
     Computer Science Department
     Williams College
     Williamstown, MA 01267
     Phone: (413) 597-2178
     FAX: (413) 597-4116

     Tom Fawcett
     Bell Atlantic Science and Technology

     Foster Provost
     Bell Atlantic Science and Technology

For additional information, see