Skip to content Skip to navigation

Colloquium: Graphical analysis of recurrent-events data with applications to engineering, medicine, marketing, and other fields

Event Type: 
Colloquium
Speaker: 
Dr. Wayne B. Nelson
Event Date: 
Friday, November 7, 2014 -
3:00pm to 4:00pm
Location: 
SMLC 356
Audience: 
General PublicFaculty/StaffStudentsAlumni/Friends
Sponsor/s: 
stats group

Event Description: 

In many applications, one has recurrent events data on sample units.
 Examples include the number and cost of repairs of products, recurrent
disease episodes in patients, sales to customers on Amazon.com, births of
babies to statisticians, etc. Analysis of such recurrence data requires
special statistical models and methods not covered in basic courses.  This
talk presents a general simple and informative model and plot for analyzing
such recurrence data on numbers or costs of recurrences.  The plots and
analyses are illustrated with data on car transmission repairs, bladder
tumor and herpes recurrences, childbirths to statisticians, amazon.com
sales, and other applications.  Computer programs that calculate and make
the plots and comparisons with confidence limits are surveyed.
The plots provide:
1.   An estimate of the average number or cost of recurrences per unit
during a period of interest, for example, warranty or design life of
products and profit on Amazon.com sales.
2.   The behavior of the population recurrence rate -- increases or
decreases with population age; this information is useful for decisions on
product burn-in, overhaul, and retirement and on patient treatments.
3.   Predictions of the future number or cost of recurrences for a unit or
the population; this is useful for predicting warranty costs and the demand
for replacement parts and for predicting patient costs for recurrent
diseases.
4.   A comparison of data sets from different populations; this is used to
decide which designs, materials, treatments, environments, etc., produce
lower product repair rates,  which promotion yields more sales, or which
treatment yields fewer recurrences.
5.   Unsought, useful information.

Event Contact

Contact Name: Yan Lu

Contact Phone: 5052772544

Contact Email: luyan@math.unm.edu