Predictive Model for Customer Segmentation using Database Marketing Techniques
5 likes1,157 views
This document describes analyzing a donor dataset to predict whether a prospect will donate or not donate. It discusses the project goals, tools used, data preprocessing steps, and various predictive models tested including CHAID, forward regression, backward regression, and stepwise regression. The models are compared on their ability to capture the top donors based on cumulative lift charts. Forward regression performed best at capturing the top 20% of donors, while backward regression captured the top 30% best. Adding an additional correlated variable to forward regression did not significantly improve performance.
Predictive Model for Customer Segmentation using Database Marketing Techniques
- 1. Database Marketing and CRM – Analyzing DONOR data set Akanksha Jain
- 2. Project Goals • Goal: Using historical data set DONOR_RAW, develop a model which can predict whether the prospect will donate/ not donate • Scope: DONOR_RAW data set • 50 Variables • 19,372 observaKons • Dependent Variable: TARGET_B(Binary) • Responder: 1 • Non-‐Responder: 0
- 3. TOOLS • SAS Enterprise Miner 4.3 • SAS 9.3_M1
- 4. Diagram
- 6. Data Source • Reject Variables: • TARGET_D (using TARGET_B as target) • ID (an id number) • WEALTH_RATING (huge no. of missing values) • Variable TARGET_B • Change Role to TARGET • Change Order to DESCENDING • Select complete data set as Sample • Set Prior ProbabiliKes • Responder: 0.05 • Non-‐Responder: 0.95
- 7. Data Partition • Train – 60% • Validate – 25% • Test – 15%
- 8. Variable Transformation Taking Log TransformaKon to reduce Skewness • LIFETIME_GIFT_RANGE • LIFETIME_MAX_GIFT_AMT • LIFETIME_MIN_GIFT_AMT • MOR_HIT_RATE • FILE_AVG_GIFT • LIFETIME_AVG_GIFT_AMT • PCT_ATTRIBUTE1 • LAST_GIFT_AMT • RECENT_AVG_GIFT_AMT Keep all variables, original and log transformaKons
- 9. Model: CHAID • Nominal Criterion: Chi Square • Significance Level: 0.1 • Minimum number of observaKons in a leaf = 25 • ObservaKons required for a split search = 55 • Model assessment measure: Total Leaf Impurity (Gini Index)
- 11. Model: CHAID (con’t) Inference: FREQUENCY_STATUS_97NK = 3 or 4; MONTHS_SINCE_LAST_GIFT < 8.5 1% = 56% Less MarkeKng Effort needed as most likely that they will donate anyways FREQUENCY_STATUS_97NK = 3 or 4; MONTHS_SINCE_LAST_GIFT >= 8.5; NUMBER_PROM_12 <11.5 1% = 43% Will also donate but the company should be careful and not send them too many promoKons FREQUENCY_STATUS_97NK = 3 or 4; MONTHS_SINCE_LAST_GIFT >= 8.5; NUMBER_PROM_12 >= 11.5 1% = 30% Are geong too many promoKons; and hence company should cut on sending them promoKons FREQUENCY_STATUS_97NK = 1, 2 or Missing 1% = 21% Study them more closely as in why they are not donaKng, what other factors are responsible and then decide how to design a markeKng campaign for them.
- 12. Variable Selection • Target AssociaKons: Select Chi Square
- 13. Model: Forward Regression MODEL OPTIONS -‐> INPUT CODING -‐>DEVIATION SELECTION METHOD -‐> FORWARD CRITERIA -‐> CROSS VALIDATION MISCLASSIFICATION ADVANCED -‐> OPTIMIZATION METHOD -‐> NEWTON-‐RAPHSON w/ LINE SEARCH • SL Entry: 0.05 • • • •
- 14. Model: Forward Regression (con’t)
- 15. Model: Forward Regression (con’t)
- 16. Model: Backward Regression MODEL OPTIONS -‐> INPUT CODING -‐>DEVIATION SELECTION METHOD -‐> BACKWARD CRITERIA -‐> CROSS VALIDATION MISCLASSIFICATION ADVANCED -‐> OPTIMIZATION METHOD -‐> NEWTON-‐RAPHSON w/ LINE SEARCH • SL Stay: 0.05 • • • •
- 17. Model: Backward Regression (con’t)
- 18. Model: Backward Regression (con’t)
- 19. Model: Stepwise Regression MODEL OPTIONS -‐> INPUT CODING -‐>DEVIATION SELECTION METHOD -‐> STEPWISE CRITERIA -‐> CROSS VALIDATION MISCLASSIFICATION ADVANCED -‐> OPTIMIZATION METHOD -‐> NEWTON-‐RAPHSON w/ LINE SEARCH • SL Entry: 0.15 • SL Stay: 0.05 • • • •
- 20. Model: Stepwise Regression (con’t)
- 21. Model: Stepwise Regression (con’t)
- 22. Variable Comparison Forward Backward Stepwise FILE_CARD_GIFT FILE_CARD_GIFT FILE_CARD_GIFT FREQUENCY_STATUS_97NK FREQUENCY_STATUS_97NK FREQUENCY_STATUS_97NK INCOME_GROUP INCOME_GROUP INCOME_GROUP LIFE_AV9* LIFE_AV9* LIFE_AV9* MONTHS_SINCE_LAST_GIFT MONTHS_SINCE_LAST_GIFT MONTHS_SINCE_LAST_GIFT PEP_STAR PEP_STAR PEP_STAR LIFETIME_GIFT_AMOUNT MEDIAN_HOUSEHOLD_INCOME RECENT_RESPONSE_PROP *LIFE_AV9 is the log(LIFETIME_AVG_GIFT_AMOUNT)
- 23. Model Comparison (Validation): Cumulative LIFT
- 24. Model Comparison (Validation): Cumulative LIFT Inference: • Capture top 20% of the market -‐>FORWARD • Capture top 30% of the market -‐>BACKWARD
- 25. Model Comparison (TEST): Cumulative LIFT
- 26. Model Comparison (TEST): Cumulative LIFT Inference: • Capture top 20% of the market -‐>FORWARD • Capture top 30% of the market -‐> ANY
- 27. Forward versus Backward • Variables: • LIFETIME_GIFT_AMOUNT • MEDIAN_HOUSEHOLD_INCOME • RECENT_RESPONSE_PROP • CorrelaKons: • MEDIAN_HOUSEHOLD_INCOME and INCOME_GROUP = 43% • LIFE_AV9 and LIFETIME_AVG_GIFT_AMT = 83% • FILE_CARD_GIFT and RECENT_RESPONSE_PROP = 30%
- 28. Model: Forward + RECENT_RESPONSE_PROP • Variable SelecKon (call it Variable_1Extra): • • • • • • • FILE_CARD_GIFT FREQUENCY_STATUS_97NK INCOME_GROUP LIFE_AV9 MONTHS_SINCE_LAST_GIFT PEP_STAR RECENT_RESPONSE_PROP • Reject other variables manually • Call this model For_1Extra
- 29. Model Comparison (Validation): Cumulative LIFT
- 30. Model Comparison (Validation): Cumulative LIFT Inference: • Capture top 20% of the market -‐>For_1Extra • Capture top 30% of the market -‐> BACKWARD
- 31. Model Comparison (TEST): Cumulative LIFT
- 32. Model Comparison (TEST): Cumulative LIFT Inference: • Capture top 20% of the market -‐>For_1Extra • Capture top 30% of the market -‐> ANY
- 33. Model: Decision Final Model: FOR_1EXTRA Variables: • • • • • • • FILE_CARD_GIFT FREQUENCY_STATUS_97NK INCOME_GROUP LIFE_AV9 MONTHS_SINCE_LAST_GIFT PEP_STAR RECENT_RESPONSE_PROP
- 34. Interaction Terms • FREQ_PEP = FREQUENCY_STATUS_97NK * PEP_STAR • FREQ_MONTH = FREQUENCY_STATUS_97NK * MONTHS_SINCE_LAST_GIFT • FREQ_INCOME = FREQUENCY_STATUS_97NK * INCOME_GROUP
- 35. Model: Forward Regression with Interaction Terms Rename model as FOR_1E_INT MODEL OPTIONS -‐> INPUT CODING -‐>DEVIATION SELECTION METHOD -‐> FORWARD CRITERIA -‐> CROSS VALIDATION MISCLASSIFICATION ADVANCED -‐> OPTIMIZATION METHOD -‐> NEWTON-‐RAPHSON w/ LINE SEARCH • SL Entry: 0.05 • • • • •
- 36. Model FOR_1E_INT: Cumulative LIFT
- 37. Model FOR_1E_INT: Variable List
- 38. Model Comparison (Validation): Cumulative LIFT
- 39. Model Comparison (Validation): Cumulative LIFT Inference: • Capture top 20% of the market -‐>FOR_1E_INT • Capture top 30% of the market -‐> FOR_1EXTRA
- 40. Model Comparison (TEST): Cumulative LIFT
- 41. Model Comparison (TEST): Cumulative LIFT Inference: • Capture top 20% of the market -‐>ANY • Capture top 30% of the market -‐> FOR_1EXTRA
- 42. Model: For_1EXTRA + Interaction terms • Variable SelecKon (call it Variable_UNION): • • • • • • • • • • FILE_CARD_GIFT FREQENCY_STATUS_97NK INCOME_GROUP LIFE_AV9 MONTHS_SINCE_LAST_GIFT PEP_STAR RECENT_RESPONSE_PROP FREQ_PEP FREQ_MONTH FREQ_INCOME • Reject other variables manually • Call this model For_Union
- 43. Model Comparison (Validation): Cumulative LIFT
- 44. Model Comparison (Validation): Cumulative LIFT Inference: • Capture top 20% of the market -‐> FOR_1E_INT • Capture top 30% of the market -‐> FOR_UNION
- 45. Model Comparison (Test): Cumulative LIFT
- 46. Model Comparison (Test): Cumulative LIFT Inference: • Capture top 20% of the market -‐> ANY • Capture top 30% of the market -‐> FOR_UNION
- 47. Model: Decision Final Model: FOR_1EXTRA because • No significant improvement with other models • InteracKon terms bring along complexity Variables: • • • • • • • FILE_CARD_GIFT FREQUENCY_STATUS_97NK INCOME_GROUP LIFE_AV9 MONTHS_SINCE_LAST_GIFT PEP_STAR RECENT_RESPONSE_PROP
- 49. THANK YOU