Using Data Mining Methods to Build Customer Profiles

  • Authors:
  • Gediminas Adomavicius;Alexander Tuzhilin

  • Affiliations:
  • -;-

  • Venue:
  • Computer
  • Year:
  • 2001

Quantified Score

Hi-index 4.11

Visualization

Abstract

Personalization--the ability to provide content and services tailored to individuals' preferences and behavior--is an important marketing tool. The authors developed an approach that uses information from customers' transactional histories to construct individual profiles containing facts about the customer and rules describing that customer's behavior. They used data mining methods to derive behavioral rules from the data. They also developed a method for validating customers' profiles to separate "good" rules from "bad." They implemented the profile construction and validation methods using the 1:1Pro system. Including personal behavioral rules in customers' profiles makes this approach unique.Focusing on statistical validity and acceptability to an expert, the authors explain how to judge the quality of rules stored in customers' profiles. Their objective is to incorporate the concept of effectiveness in 1:1Pro. The authors propose combining the constraint specification, data mining, and rule validation stages into one system.