Customer Retention via Data Mining

摘要

``Customer Retention'' is an increasingly pressing issue intoday's ever-competitive commercial arena. This is especially relevantand important for sales and services related industries. Motivated by areal-world problem faced by a large company, we proposed a solution thatintegrates various techniques of data mining, such as featureselection via induction, deviation analysis, and mining multipleconcept-level association rules to form an intuitive and novel approachto gauging customer loyalty and predicting their likelihood ofdefection. Immediate action triggered by these ``early-warnings''resulting from data mining is often the key to eventual customerretention.