Author of “Data Mining Techniques: For Marketing, Sales, and Customer Relationship Management” which came out in 2004, and updated in 2011.

Abstract: Survival analysis, also called time-to-event analysis, is an  underutilized tool in the data miner's toolkit. Because many analytical  practitioners in the business world are unfamiliar with survival  analysis, they have become adept at reformulating questions that should  properly be about "when" into questions about "whether." A typical  attrition model asks "Will the customer still be active in six months?"  This is a convenient question because it has the kind of yes or no  answer that can easily be modeled using logistic regression, or decision  trees, but a much more interesting question is "How long will the  customer last?" The answer to that question has many uses from lifetime  value calculations to optimizing acquisition channels. The particular  application addressed in this talk is creating a long-range daily  forecast for a subscriber population that is a constantly changing mix  of customers segments, each with its own survival function. A bottom-up  forecast of this sort is more useful as a planning tool than the typical  aggregate forecast because the survival-based forecast is very  sensitive to changes in assumptions about the characteristics of new  customers. The analyst can study the effect of varying assumptions about  credit scores, product preferences, rate plans, demographics, or  whatever else is of interest  by watching the effect of these  assumptions on the forecast. The work described here is based on several  successful forecasts developed for newspapers, telephone companies, and  other subscription-based businesses.

Location: First Floor of the Maureen Murphy Wilkens Science Center at Emmanuel College,WSC 102
Directions: http://www.emmanuel.edu/About_Emmanuel/Campus/Wilkens_Science_Center/FAQs.html

Sponsored by the Boston Chapters of the American Statistical Association and we thank Emmanuel College for hosting.

Time: 6:15 PM light dinner, 7:00 PM lecture

Directions: http://www.emmanuel.edu/Tools_Navigation/Maps_and_Directions.html

To register by check, include the check made payable to BCASA, your name, affiliation, and mail by March 19 to:

Huichao Chen, PhD 
Department of Biostatistics/CBAR 
Harvard University 
651 Huntington Ave, FXB502 
Boston, MA 02115