Two New Directions for Data Mining Charles Ling

Two New Directions for Data Mining Charles Ling, Ph. D Department of Computer Science University of Western Ontario, Canada Director, Data Mining Lab, UWO cling@csd. uwo. ca http: //csd. uwo. ca/faculty/cling Charles Ling, Ph. D

Two New Directions for Data Mining l Action Mining l Active Cost-sensitive Learning Charles Ling, Ph. D

Action Mining for Profitable CRM Charles Ling, Ph. D Department of Computer Science University of Western Ontario, Canada Director, Data Mining Lab, UWO cling@csd. uwo. ca http: //csd. uwo. ca/faculty/cling Charles Ling, Ph. D

CRM Customer Relationship Management: focus on customer satisfaction to improve profit Two kinds of CRM l Enabling CRM: Infrastructure, multiple touch point, data integration and management, … – Oracle, IBM, People. Soft, Siebel Systems, … l Intelligent CRM: data – Vendors/products mining and data analysis (http: //www. kdnuggets. com/solutions/crm. html) Charles Ling, Ph. D

Three Intelligent CRM Tasks l Acquisition: direct marketing, application form, promotion methods, … l Customization: cross/up-sale, segmentation, promotions, … l Retention: Attrition/churn prevention Goal: through data mining to improve customer loyalty, satisfaction, and spending, resulting in increased company profits Charles Ling, Ph. D

Action Mining Beyond model building and customer profiling Improve customer relationship: Actions changes What actions should you take to change customers from an undesired status to a desired one – – – From churn to loyal From inactive to active From low spending to high spending From non-customers to customers … and make the maximum profit (the ultimate goal) Charles Ling, Ph. D

Research Issues Bounded Action Problem (BAP) – Types of actions are limited to k – How to find k action types to maximize profit l The problems are NP-hard – Exponential to k l Our solutions: heuristic/greedy search based on decision trees – Proactive Solution Charles Ling, Ph. D

How Proactive Solution Works 1. Get Customer Data (marketing DB) 2. Build Customer Profiles 3. Search Actions for Maximal Profit 4. Action Delivery Charles Ling, Ph. D

Step 1: Get Customer Data Marketing DB: Segmentation, data preparation, pre-processing… Define a “target”: undesired status and desired status ID Name Age Sex Service Rate Prof … Retained (Target) 1001 John 50 M H L A … Yes 3010 Sue 25 F M H D … No … … … … 40 M M H B … ? ? ? … 1112 Jack Charles Ling, Ph. D

Step 2: Build Customer Profile on target Automatically by Proactive Solution with probabilities on the target Service M H L Sex F Rate M L H Prob=0. 1 Prob=0. 9 Charles Ling, Ph. D Prob = 0. 2 Prob=0. 8 Prob=0. 5

Step 3: Search Actions for Maximal Profit Proactive Solution searches more desired nodes in the profile… ID Name Age Sex Service Rate Prof … Retained … … … … … 40 M M H B … ? ? ? 1112 Jack Charles Ling, Ph. D

Jack: …, Service = M, Sex = M, Rate = H, … Profit =$4000 Service M H L Sex F Rate M L Prob=0. 1 H Prob M H Serv: gain = -0. 1 E. Profit= -400 Rate: $500 L Cost= H E. Net Profit= -900 Prob=0. 9 Prob gain = 0. 7 E Profit= $2800 Cost = E Net Profit= - Charles Ling, Ph. D Prob = 0. 2 Prob=0. 8 Prob gain = 0. 6 E Profit=$2400 E. Profit=$2400 Cost=$800 E Net. Profit=$1600 E. Net. Profit=$1600 Prob=0. 5 Prob gain = 0. 3 E Profit=$1200 Cost=$400 E Net. Profit=$800

Step 4: Action Deployment ID Name Prob Actions diff 1112 Jack … 0. 6 3010 Sue 0. 5 3421 Bill … Action costs Net. Profit Service: M H $800 Rate: H L … $1600 Sig. Acc: 0 1 Service: L M … $700 N/A $500 $0 • Selective deployment: human intelligence, … • Customer segmentation by actions Charles Ling, Ph. D

Reporting – on the web Charles Ling, Ph. D

Advanced Features l Accurate probability estimations l Better evaluation methods – AUC of ROC l Hard vs soft attributes – search many trees l Beam-search l Action correlation Charles Ling, Ph. D

Case Study: Mutual Fund An insurance company selling mutual funds l Task 1: For the current fund owners, how to improve their fund purchasing (from low to high spending)? l Task 2: Some representatives are good performers but some are not; how to change bad performers to be good performers? l Task 3: Many customers do not currently own mutual funds. How to market to them to buy mutual funds? Charles Ling, Ph. D

Summary l From model building to action mining (deployment) l Business oriented: maximal net profit l Proactive Solution: effective intelligent CRM l Technically sophisticated l Massive one-to-one customization l Effective marketing and segmentation tool Charles Ling, Ph. D