SUPPORT VECTOR MACHINES FOR DIFFERENTIAL PREDICTION Finn Kuusisto 1 - - PowerPoint PPT Presentation

β–Ά
support vector machines
SMART_READER_LITE
LIVE PREVIEW

SUPPORT VECTOR MACHINES FOR DIFFERENTIAL PREDICTION Finn Kuusisto 1 - - PowerPoint PPT Presentation

Nov 09, 2023 362 likes β€’609 views

SUPPORT VECTOR MACHINES FOR DIFFERENTIAL PREDICTION Finn Kuusisto 1 , Vitor Santos Costa 2 , Houssam Nassif 3 , Elizabeth Burnside 1 , David Page 1 , and Jude Shavlik 1 1 University of Wisconsin Madison 2 University of Porto 3 Amazon 9/18/2014

slide-1
SLIDE 1

Finn Kuusisto 1, Vitor Santos Costa 2, Houssam Nassif3, Elizabeth Burnside 1, David Page 1, and Jude Shavlik 1

9/18/2014 Support Vector Machines for Differential Prediction

SUPPORT VECTOR MACHINES FOR DIFFERENTIAL PREDICTION

1University of Wisconsin – Madison 2University of Porto 3Amazon

slide-2
SLIDE 2

9/18/2014 Support Vector Machines for Differential Prediction

DIFFERENTIAL PREDICTION

Goal

Use modeling techniques to gain insight about the differences between two subgroups of a population.

slide-3
SLIDE 3

9/18/2014 Support Vector Machines for Differential Prediction

UPLIFT MODELING

( R A D C L I F F E & S I M P S O N , 2 0 0 8 )

How do we choose which customers to target with some marketing activity?

slide-4
SLIDE 4

9/18/2014 Support Vector Machines for Differential Prediction

UPLIFT MODELING

( R A D C L I F F E & S I M P S O N , 2 0 0 8 )

Persuadables Customers who respond positively to marketing activity. Sure Things Customers who respond positively regardless. Lost Causes Customers who respond negatively regardless. Sleeping Dogs Customers who respond negatively to marketing activity.

How do we choose which customers to target with some marketing activity?

slide-5
SLIDE 5

Target Control Response No Response Response No Response Persuadables, Sure Things Sleeping Dogs, Lost Causes Sleeping Dogs, Sure Things Persuadables, Lost Causes

9/18/2014 Support Vector Machines for Differential Prediction

UPLIFT MODELING

( R A D C L I F F E & S I M P S O N , 2 0 0 8 )

True customer groups are unknown.

slide-6
SLIDE 6

9/18/2014 Support Vector Machines for Differential Prediction

UPLIFT MODELING

𝑩𝑽𝑽 = 𝑩𝑽𝑴𝑼 βˆ’ 𝑩𝑽𝑴𝑫

Lift

The number of true positives that a classifier achieves at a given proportion of the population labeled positive.

Uplift

The difference in lift produced by a classifier between target and control subgroups.

slide-7
SLIDE 7

ο‚‘ Non-steroidal anti-inflammatory drug (NSAID) ο‚‘ Significantly reduced occurrence of adverse gastrointestinal effects common to other NSAIDs (e.g. ibuprofen) ο‚‘ Rapid and widespread acceptance for treatment of ailments such as arthritis ο‚‘ Later clinical trials showed increased risk of myocardial infarction (MI), or β€œheart attack”

9/18/2014 Support Vector Machines for Differential Prediction

TASK: ADVERSE COX-2 INHIBITOR EFFECTS

Identify patients who are susceptible to an increased risk of MI as a direct result of taking COX-2 inhibitors.

slide-8
SLIDE 8

9/18/2014 Support Vector Machines for Differential Prediction

UPLIFT MODELING TO MEDICINE: COX-2 INHIBITORS

Want

Identify patients who demonstrate an increased risk of MI as a direct result of being treated with COX-2 inhibitors.

Main Assumption

Patients with an increased risk of MI due to treatment with COX-2 inhibitors are directly analogous to customers with an increased chance of buying due to targeting – the persuadables.

slide-9
SLIDE 9

ο‚‘ Compared SVMUpl against 4 alternate SVM methods (2 shown) ο‚‘ 10-fold cross-validation for evaluation ο‚‘ Cost parameters selected from 10 through 10βˆ’6 ο‚‘ Mann-Whitney test at 95% confidence for per-fold AUU comparison

9/18/2014 Support Vector Machines for Differential Prediction

METHODS

slide-10
SLIDE 10

Model AUU COX-2 AUL No COX-2 AUL SVMUpl 𝒒-value π‘‡π‘Šπ‘π‘‰π‘žπ‘š 50.7 123.4 72.7

  • COX-2-Only

13.8 151.5 137.7 0.002 * Standard 1.2 147.7 146.5 0.002 * Baseline 0.0 0.0 0.0 0.002 *

9/18/2014 Support Vector Machines for Differential Prediction

RESULTS: COX-2 INHIBITORS

slide-11
SLIDE 11

9/18/2014 Support Vector Machines for Differential Prediction

RESULTS: COX-2 INHIBITORS

slide-12
SLIDE 12

Extend previous SVM work maximizing AUC (Joachims, 2005) to maximize AUU instead.

9/18/2014 Support Vector Machines for Differential Prediction

HOW

slide-13
SLIDE 13

9/18/2014 Support Vector Machines for Differential Prediction

SVM FOR UPLIFT

𝐡𝑉𝑀 = 𝑄 Γ— 𝜌 2 + 1 βˆ’ 𝜌 𝐡𝑉𝐷 Let the positive skew of data be: 𝜌 = 𝑄 𝑄 + 𝑂 Then (Tuffery, 2011):

slide-14
SLIDE 14

9/18/2014 Support Vector Machines for Differential Prediction

SVM FOR UPLIFT

𝐡𝑉𝑉 = π΅π‘‰π‘€π‘ˆ βˆ’ 𝐡𝑉𝑀𝐷 = π‘„π‘ˆ Γ— πœŒπ‘ˆ 2 + 1 βˆ’ πœŒπ‘ˆ π΅π‘‰π·π‘ˆ βˆ’ 𝑄𝐷 Γ— 𝜌𝐷 2 + 1 βˆ’ 𝜌𝐷 𝐡𝑉𝐷𝐷 𝑛𝑏𝑦 𝐡𝑉𝑉 ≑ 𝑛𝑏𝑦 π΅π‘‰π·π‘ˆ βˆ’ πœ‡π΅π‘‰π·π· 𝑛𝑏𝑦 𝐡𝑉𝑉 ≑ 𝑛𝑏𝑦 π‘„π‘ˆ Γ— 1 βˆ’ πœŒπ‘ˆ π΅π‘‰π·π‘ˆ βˆ’ 𝑄𝐷 Γ— 1 βˆ’ 𝜌𝐷 𝐡𝑉𝐷𝐷 ∝ 𝑛𝑏𝑦 π΅π‘‰π·π‘ˆ βˆ’ 𝑄𝐷 Γ— 1 βˆ’ 𝜌𝐷 π‘„π‘ˆ Γ— 1 βˆ’ πœŒπ‘ˆ 𝐡𝑉𝐷𝐷 πœ‡

slide-15
SLIDE 15

ο‚‘ Most common cancer in women ο‚‘ Two basic stages: In situ and invasive

  • In situ cancer cells are localized
  • Invasive cancer cells have infiltrated surrounding tissue

ο‚‘ Younger women tend to have more aggressive in situ cancer ο‚‘ Older women sometimes have indolent in situ cancer

9/18/2014 Support Vector Machines for Differential Prediction

TASK: IN SITU BREAST CANCER

Identify older patients with indolent in situ breast cancer.

slide-16
SLIDE 16

Want

Identify older patients with in situ breast cancer that is distinct from that of younger patients.

Main Assumption

Older patients with in situ breast cancer that is distinct from that of younger patients, who tend to have aggressive cancer, have a decreased risk of invasive progression.

9/18/2014 Support Vector Machines for Differential Prediction

UPLIFT MODELING TO MEDICINE: BREAST CANCER

slide-17
SLIDE 17

Model AUU Older AUL Younger AUL SVMUpl 𝒒-value π‘‡π‘Šπ‘π‘‰π‘žπ‘š 19.2 64.3 45.1

  • Older-Only

5.9 67.7 61.9 0.037 * Standard 11.0 75.4 64.3 0.049 * Baseline 11.0 66.0 55.0 0.004 *

9/18/2014 Support Vector Machines for Differential Prediction

RESULTS: BREAST CANCER

slide-18
SLIDE 18

9/18/2014 Support Vector Machines for Differential Prediction

RESULTS: BREAST CANCER

slide-19
SLIDE 19

9/18/2014 Support Vector Machines for Differential Prediction

UPLIFT MODELING SIMULATION

ο‚‘ Generated synthetic customer population ο‚‘ Subjected customer population randomly to simulated marketing activity ο‚‘ Measured uplift as usual ο‚‘ Measured ROC with Persuadables as the positive class, others as negative

slide-20
SLIDE 20

9/18/2014 Support Vector Machines for Differential Prediction

UPLIFT MODELING SIMULATION: UPLIFT CURVE

slide-21
SLIDE 21

9/18/2014 Support Vector Machines for Differential Prediction

UPLIFT MODELING SIMULATION: PERSUADABLE ROC

slide-22
SLIDE 22

ο‚‘ Extended previous SVM work on AUC maximization to AUU ο‚‘ Results suggest SVMUpl achieves better uplift than many alternate SVM methods ο‚‘ May want to make performance guarantees for control group ο‚‘ May want to interpret learned model ο‚‘ Better verification that maximizing uplift is appropriate goal

9/18/2014 Support Vector Machines for Differential Prediction

CONCLUSIONS & FUTURE WORK

slide-23
SLIDE 23

Questions?

9/18/2014 Support Vector Machines for Differential Prediction

THANKS

slide-24
SLIDE 24

Radcliffe, N. and Simpson. R.: Identifying who can be saved and who will be driven away by retention activity. Journal of Telecommunications Management (2008). Tuffery, S.: Data Mining and Statistics for Decision Making. John Wiley & Sons, 2nd edn. (2011). Joachims, T.: A support vector method for multivariate performance

  • measuers. In: Proceedings of the 22 nd International Conference on

Machine Learning (2005).

9/18/2014 Support Vector Machines for Differential Prediction

SELECTED REFERENCES