UpliftEvaluation_NRA

You can find the code used to 1- generate biased samples 2- Treatment group reweighting with the class transformation approach, the 2 Model approach, S-Learner

To execute the bias generation process with a dataset:

1- The treatment and output columns should be binary (0 and 1) and named respectively : 'segment' and 'visit'
2- Datasets variables should be discretized

You can launch the process with the following command on a Linux machine:

(For the bias generation process without domain adaptation for samples reweighting):

ipython main.py DatasetFileName UpliftMethod X NoDA VAR_Name

(For the bias generation process with the reweighting method "Ratio of gaussians"):

ipython main.py DatasetFileName UpliftMethod X rg VAR_Name

For the Uplift methods, you can use :

1- KL [2]
2- CTS [3]
3- Chi [2]
4- ED [2]
5- 2M_LR (Two Model Approach with Logistic Regression)[4]
6- 2M_Xgboost (Two Model Approach with Xgboost) [4]
7- CT_Xgboost (Class Transformation Approach with Xgboost) [5]
8- CT_LR (Class Transformation with Logistic Regression) [5]
9- DR_LinR Doubly Robust Learner with Linear regression
10- DR_Xgboost Doubly Robust Learner with Xgboost
11- XLearner_LinR XLearner with Linear regression
12- XLearner_Xgboost XLearner with Xgboost
13- SLearner_LinR SLearner with Linear regression
14- SLearner_Xgboost SLearner with Xgboost

For the VAR_Name you can use the variable name as it appears in the data, or to use "Comb" in order to generate bias using all the variables in the dataset

Datasets

In our study we used this code with the following datasets:

Hillstrom [6]: Marketing dataset with 64.000 lines in total. It contains two groups of treatments and a control group. The treatments are different marketing campaigns sent by mail, including advertisements of women's products (21.387 lines) or men's products (21.307 lines). The control group contains the individuals who did not receive a marketing campaign. A positive response is considered when the customer visits the website within two weeks after receiving the treatment (the e-mail). In the experiments, we used the treatment of the women's products campaign and the control group.
Criteo [1]: This database was created specifically for uplift prediction. It contains 14 million individuals, where one group of individuals was targeted by an advertising campaign and another randomly selected group that was not targeted by the campaign. The treatment rate is 85%, and a visit from the customer marks a positive response. A sample of 50.000 individuals was used.
Gerber [10]: A database where we can study political behavior and the effect of social pressure to vote. Mailings were sent randomly to voters to voters 11 days before an election in 2006. Voters who received Neighbors' messages constitute the treatment group. In contrast, individuals who received the 'Civic Duty' message are the control group. Positive Response is considered when the voter votes.
Retail Hero [7]: X5 sales group dataset that sends SMS messages to encourage people to buy more. The treatment group is the group of individuals who have been contacted by message, and the control group is the group that was not contacted. A positive response is considered when the customer has made a purchase after being contacted.
Zenodo [8]: a synthetic database containing trigonometric patterns specifically designed for uplift evaluation. We used a subset of 20,000 rows of data (data identified by the variable trial_id = 1 and trial_id = 2).
Megafon [9]: Synthetic database was created specifically for uplift prediction. Telecom companies generated it in order to reproduce the situations encountered them.
Synth1 and Synth2 : (Provided in the repo.) 2 synthetic datasets that we have built as a 2D grid of size 10x10 in which each cell corresponds to a particular uplift drawn at random.

Results

Following are the results of the benchmark that demonstrate the behavior of several uplift methods when facing NRA bias on the datasets cited above :

{}	2M_Xgboost	2M_LR	CT_Xgboost	CT_LR	DR_Xgboost	DR_LinR	XLearner_Xgboost	XLearner_LinR	SLearner_LinR	SLearner_Xgboost	KL	ED	CTS
Criteo_f2	6.6(1.7)	7.2(1.6)	0.2(1.9)	1.9(1.2)	4.4(2.8)	9.9(0.9)	5.5(2.6)	8.5(0.8)	-0.2(1.9)	8.0(1.9)	0.6(1.4)	4.9(1.3)	2.1(1.5)
Criteo_f2'	6.4(2.7)	10.1(1.4)	6.9(2.0)	7.6(1.4)	4.4(2.7)	10.3(1.6)	4.1(2.1)	11.1(1.0)	-0.2(2.1)	6.2(2.7)	10.3(1.2)	11.0(1.2)	10.5(1.3)
Criteo_f8	8.1(2.6)	6.3(2.0)	0.1(1.7)	1.7(1.0)	3.7(2.3)	9.8(1.0)	5.4(2.6)	8.1(1.1)	-0.2(1.7)	8.4(1.9)	1.2(1.6)	5.2(1.2)	2.4(1.6)
Criteo_f8'	7.2(2.6)	7.1(2.2)	-2.5(1.7)	-2.0(1.3)	3.4(3.0)	10.3(0.9)	5.3(2.0)	9.1(1.1)	-0.4(1.9)	6.9(2.1)	-1.3(1.5)	4.5(1.6)	2.3(1.9)
Gerber_p2002	-2.4(2.0)	1.1(1.1)	-2.1(1.5)	-0.4(1.2)	-2.0(1.9)	0.8(1.1)	-2.3(1.9)	1.4(1.1)	0.1(0.9)	-2.0(2.0)	-1.5(1.8)	-0.9(1.5)	-0.1(1.7)
Gerber_p2002'	-2.4(1.9)	0.6(1.1)	-1.5(2.2)	1.0(1.1)	-1.7(1.6)	0.8(1.2)	-2.2(1.4)	0.7(1.0)	-0.2(0.9)	-1.9(2.2)	-0.8(1.6)	-0.4(1.6)	0.1(1.6)
Gerber_p2004	-2.1(2.0)	0.8(1.1)	-1.8(1.7)	-1.2(1.3)	-2.1(1.9)	0.7(1.1)	-2.1(1.8)	1.2(1.3)	0.0(1.1)	-1.8(2.0)	-1.7(1.8)	-1.5(1.9)	-0.6(1.9)
Gerber_p2004'	-2.4(2.0)	1.0(1.3)	-1.1(2.4)	1.4(1.4)	-2.1(1.7)	1.1(1.5)	-2.0(1.7)	0.9(1.0)	0.1(1.1)	-1.9(2.0)	-1.3(1.3)	-0.9(1.4)	0.0(1.5)
Hillstrom_mens	2.7(2.1)	5.5(2.6)	-4.1(2.0)	-4.6(2.2)	1.9(2.4)	5.4(2.1)	2.0(2.6)	5.5(2.2)	0.2(2.4)	2.5(2.7)	2.8(2.6)	2.9(2.5)	1.0(2.8)
Hillstrom_mens'	2.7(2.5)	6.5(2.7)	4.6(1.5)	6.6(1.9)	1.2(2.2)	5.9(2.4)	1.8(2.5)	5.8(1.8)	0.0(2.2)	2.9(2.9)	4.8(2.5)	4.9(2.1)	6.0(2.1)
Hillstrom_newbie	2.8(2.2)	6.2(2.7)	0.1(2.1)	2.4(1.9)	1.0(2.4)	5.9(2.0)	2.1(2.3)	6.0(2.0)	-0.1(2.4)	3.3(2.2)	4.2(2.2)	4.3(2.5)	4.3(2.5)
Hillstrom_newbie'	3.6(2.6)	6.2(2.2)	1.7(2.1)	4.1(1.9)	1.8(2.7)	6.0(1.7)	2.9(2.8)	6.3(2.2)	0.2(2.4)	2.5(2.7)	4.7(1.9)	4.8(2.0)	4.9(2.3)
Megafone_X16	17.8(0.5)	3.5(0.4)	8.6(0.6)	3.2(0.4)	16.9(0.5)	3.0(0.5)	18.3(0.4)	3.0(0.6)	-0.0(0.6)	17.9(0.4)	13.2(0.5)	13.7(0.5)	11.6(0.7)
Megafone_X16'	18.0(0.4)	3.6(0.4)	14.2(0.5)	3.1(0.5)	17.3(0.5)	3.3(0.4)	18.4(0.5)	3.4(0.4)	-0.1(0.5)	18.2(0.5)	13.5(0.5)	13.7(0.5)	12.2(0.6)
Megafone_X21	18.2(0.4)	3.5(0.4)	12.0(0.4)	2.4(0.5)	17.4(0.5)	3.0(0.4)	18.8(0.4)	3.1(0.4)	-0.0(0.6)	18.4(0.4)	13.9(0.5)	14.0(0.6)	10.7(0.8)
Megafone_X21'	18.2(0.5)	3.3(0.6)	12.8(0.4)	1.8(0.6)	16.9(0.5)	2.7(0.6)	18.4(0.4)	2.9(0.5)	0.0(0.5)	18.3(0.4)	13.6(0.5)	13.9(0.5)	11.9(0.6)
Synth1	7.0(0.9)	0.9(1.6)	1.7(0.9)	-2.9(1.3)	9.7(1.5)	-0.4(1.5)	12.6(1.6)	-1.6(2.0)	0.6(1.6)	12.2(1.2)	9.7(1.2)	8.8(1.6)	8.7(1.2)
Synth1'	9.5(1.5)	2.1(0.7)	-0.4(3.0)	-0.8(2.8)	13.1(1.1)	0.3(1.9)	9.4(1.5)	-0.8(1.0)	-1.0(1.6)	9.9(1.4)	12.0(1.9)	11.1(2.3)	10.4(1.8)
Synth2	9.8(0.1)	1.9(0.1)	8.1(0.5)	1.1(0.2)	9.7(0.2)	1.9(0.1)	9.7(0.2)	1.8(0.1)	-0.1(0.4)	10.1(0.1)	9.7(0.1)	9.6(0.2)	8.7(0.1)
Synth2'	10.0(0.1)	1.7(0.0)	8.6(0.1)	1.4(0.4)	9.9(0.2)	1.7(0.1)	9.8(0.1)	1.7(0.1)	0.1(0.5)	9.9(0.2)	9.8(0.1)	9.5(0.2)	8.8(0.2)
retailHero_age	0.7(0.4)	1.2(0.3)	0.3(0.4)	0.8(0.4)	0.5(0.4)	1.3(0.4)	0.5(0.3)	1.2(0.3)	-0.0(0.3)	0.9(0.3)	0.8(0.3)	0.9(0.3)	0.9(0.4)
retailHero_age'	0.7(0.5)	1.2(0.3)	0.6(0.4)	1.3(0.4)	0.3(0.4)	1.2(0.4)	0.6(0.4)	1.2(0.4)	-0.0(0.4)	0.8(0.4)	0.9(0.4)	0.9(0.4)	0.6(0.5)
retailHero_trNum	0.8(0.4)	1.2(0.3)	0.4(0.3)	1.1(0.4)	0.4(0.4)	1.3(0.4)	0.5(0.4)	1.2(0.4)	-0.0(0.4)	0.9(0.4)	0.7(0.4)	0.7(0.4)	0.6(0.4)
retailHero_trNum'	0.8(0.4)	1.2(0.4)	0.5(0.4)	1.1(0.4)	0.4(0.3)	1.2(0.4)	0.4(0.3)	1.2(0.3)	0.0(0.3)	0.6(0.4)	0.6(0.4)	0.6(0.5)	0.3(0.4)
zenodoSynth_X10	9.7(1.8)	12.6(1.9)	7.0(2.2)	12.1(1.5)	7.8(1.9)	12.2(1.9)	9.4(1.7)	12.1(1.7)	0.0(2.5)	11.5(2.0)	12.8(1.9)	13.0(1.9)	10.6(2.6)
zenodoSynth_X10'	9.5(2.1)	12.6(1.8)	6.9(2.1)	12.0(1.7)	8.0(2.2)	12.3(1.8)	8.8(2.4)	12.3(2.0)	0.1(2.2)	11.8(2.1)	13.5(2.4)	13.5(2.4)	10.7(2.5)
zenodoSynth_X31	9.8(2.4)	12.2(2.0)	6.6(2.0)	12.0(1.9)	7.7(2.1)	12.3(1.9)	9.7(2.2)	12.4(1.7)	0.1(1.9)	11.7(2.2)	12.7(1.9)	13.2(2.0)	10.2(2.2)
zenodoSynth_X31'	9.5(2.4)	12.2(1.8)	9.2(2.0)	11.8(1.5)	7.5(2.1)	12.4(1.9)	9.0(2.0)	12.4(1.8)	0.2(2.1)	11.1(2.1)	13.0(2.5)	13.4(2.3)	12.1(2.1)

Following are the results of the treatment samples reweighting

The reweighted methods are marked in green and orange while the method w/o reweighting in blue

Class Transformation with LR

{}	CT_LR	CT_LR_wt1	CT_LR_wt2
Criteo_f2	1.9(1.2)	6.1(1.5)	8.2(2.0)
Criteo_f2'	7.6(1.4)	7.7(2.9)	7.2(2.1)
Criteo_f8	1.7(1.0)	5.5(1.8)	**7.9(1.7)
Criteo_f8'	-2.0(1.3)	6.1(1.4)	7.3(2.0)
Gerber_p2002	-0.4(1.2)	0.9(1.1)	0.5(1.2)
Gerber_p2002'	1.0(1.1)	**1.2(1.2)	0.6(1.3)
Gerber_p2004	-1.2(1.3)	0.9(1.3)	0.4(1.1)
Gerber_p2004'	1.4(1.4)	0.7(1.1)	0.7(1.2)
Hillstrom_mens	-4.6(2.2)	5.3(2.2)	4.2(2.2)
Hillstrom_mens'	6.6(1.9)	5.9(2.3)	6.3(1.9)
Hillstrom_newbie	2.4(1.9)	5.6(2.0)	5.2(2.1)
Hillstrom_newbie'	4.1(1.9)	6.0(1.9)	5.8(2.4)
Megafone_X16	3.2(0.4)	3.1(0.4)	3.2(0.4)
Megafone_X16'	3.1(0.5)	3.4(0.3)	3.1(0.4)
Megafone_X21	2.4(0.5)	3.1(0.4)	3.0(0.5)
Megafone_X21'	1.8(0.6)	2.5(0.5)	2.9(0.5)
Synth1	-2.9(1.3)	-1.0(1.8)	-0.8(0.9)
Synth1'	-0.8(2.8)	-0.7(1.9)	-1.3(2.0)
Synth2	1.1(0.2)	1.9(0.1)	1.7(0.1)
Synth2'	1.4(0.4)	1.6(0.3)	1.8(0.1)
retailHero_age	0.8(0.4)	1.3(0.4)	1.2(0.3)
retailHero_age'	1.3(0.4)	1.2(0.4)	1.2(0.4)
retailHero_trNum	1.1(0.4)	1.2(0.4)	1.2(0.4)
retailHero_trNum'	1.1(0.4)	1.1(0.3)	1.2(0.4)
zenodoSynth_X10	12.1(1.5)	11.9(1.7)	9.8(1.8)
zenodoSynth_X10'	12.0(1.7)	11.6(1.9)	9.9(2.0)
zenodoSynth_X31	12.0(1.9)	12.1(1.7)	9.9(2.0)
zenodoSynth_X31'	11.8(1.5)	12.1(1.8)	11.1(1.6)

Class Transformation with Xgboost

{}	CT_Xgboost	CT_Xgboost_wt1	CT_Xgboost_wt2
Criteo_f2	0.2(1.9)	2.6(1.8)	4.8(1.9)
Criteo_f2'	6.9(2.0)	5.1(1.7)	4.4(2.6)
Criteo_f8	0.1(1.7)	3.4(1.5)	5.0(1.8)
Criteo_f8'	-2.5(1.7)	1.7(2.2)	2.5(2.5)
Gerber_p2002	-2.1(1.5)	-1.6(1.9)	-2.3(1.8)
Gerber_p2002'	-1.5(2.2)	-1.6(2.0)	-2.1(1.8)
Gerber_p2004	-1.8(1.7)	-1.7(1.8)	-2.3(1.9)
Gerber_p2004'	-1.1(2.4)	-1.7(2.0)	-2.1(2.1)
Hillstrom_mens	-4.1(2.0)	-0.2(2.7)	0.5(2.4)
Hillstrom_mens'	4.6(1.5)	1.8(2.1)	2.9(2.1)
Hillstrom_newbie	0.1(2.1)	1.3(2.0)	1.4(2.1)
Hillstrom_newbie'	1.7(2.1)	1.5(2.2)	3.0(2.4)
Megafone_X16	8.6(0.6)	8.4(0.5)	15.5(0.5)
Megafone_X16'	14.2(0.5)	14.3(0.4)	16.5(0.5)
Megafone_X21	12.0(0.4)	12.0(0.4)	16.0(0.5)
Megafone_X21'	12.8(0.4)	12.7(0.4)	16.1(0.5)
Synth1	1.7(0.9)	2.5(0.7)	8.9(2.9)
Synth1'	-0.4(3.0)	0.8(2.7)	9.1(1.7)
Synth2	8.1(0.5)	8.3(0.5)	9.7(0.2)
Synth2'	8.6(0.1)	8.4(0.3)	9.7(0.1)
retailHero_age	0.3(0.4)	0.3(0.4)	0.6(0.4)
retailHero_age'	0.6(0.4)	0.7(0.4)	0.6(0.4)
retailHero_trNum	0.4(0.3)	0.4(0.3)	0.6(0.3)
retailHero_trNum'	0.5(0.4)	0.6(0.4)	0.5(0.4)
zenodoSynth_X10	7.0(2.2)	7.4(2.0)	6.5(2.1)
zenodoSynth_X10'	6.9(2.1)	7.3(2.1)	6.6(1.8)
zenodoSynth_X31	6.6(2.0)	7.2(2.5)	6.5(2.2)
zenodoSynth_X31'	9.2(2.0)	7.9(2.3)	7.4(2.1)

2 Model approach with Xgboost

{}	2M_Xgboost	2M_Xgboost_wt1	2M_Xgboost_wt2
Criteo_f2	6.6(1.7)	7.9(2.6)	7.7(1.7)
Criteo_f2'	6.4(2.7)	6.9(2.2)	7.3(2.7)
Criteo_f8	8.1(2.6)	7.3(1.7)	6.8(1.8)
Criteo_f8'	7.2(2.6)	7.2(2.2)	6.5(2.2)
Gerber_p2002	-2.4(2.0)	-2.1(1.9)	-1.8(2.0)
Gerber_p2002'	-2.4(1.9)	-2.7(1.8)	-2.3(2.5)
Gerber_p2004	-2.1(2.0)	-2.3(2.1)	-2.5(2.3)
Gerber_p2004'	-2.4(2.0)	-2.7(1.5)	-2.6(1.8)
Hillstrom_mens	2.7(2.1)	2.6(2.5)	3.5(2.0)
Hillstrom_mens'	2.7(2.5)	3.3(2.8)	2.6(2.5)
Hillstrom_newbie	2.8(2.2)	3.5(1.9)	3.1(2.6)
Hillstrom_newbie'	3.6(2.6)	3.1(2.7)	3.2(2.2)
Megafone_X16	17.8(0.5)	17.9(0.4)	17.6(0.4)
Megafone_X16'	18.0(0.4)	18.0(0.5)	17.7(0.5)
Megafone_X21	18.2(0.4)	18.2(0.6)	18.0(0.5)
Megafone_X21'	18.2(0.5)	18.1(0.5)	17.9(0.4)
Synth1	7.0(0.9)	10.9(0.7)	10.8(3.4)
Synth1'	9.5(1.5)	9.8(0.9)	10.1(1.6)
Synth2	9.8(0.1)	9.8(0.1)	9.9(0.1)
Synth2'	10.0(0.1)	10.1(0.1)	9.9(0.1)
retailHero_age	0.7(0.4)	0.7(0.3)	**0.7(0.4)
retailHero_age'	0.7(0.5)	0.7(0.4)	0.6(0.4)
retailHero_trNum	0.8(0.4)	0.8(0.4)	0.7(0.3)
retailHero_trNum'	0.8(0.4)	0.7(0.4)	0.8(0.4)
zenodoSynth_X10	9.7(1.8)	9.8(2.4)	10.5(2.3)
zenodoSynth_X10'	9.5(2.1)	9.7(2.8)	10.3(1.9)
zenodoSynth_X31	9.8(2.4)	10.3(2.0)	9.1(2.4)
zenodoSynth_X31'	9.5(2.4)	8.9(2.1)	9.5(1.9)

2 model approach with Logistic regression

{}	2MLR	2MLR_trtProb	2MLR_trtProbXgboost
Criteo_f2	7.2(1.6)	9.3(1.5)	9.0(2.0)
Criteo_f2'	10.1(1.4)	9.7(1.4)	10.4(1.8)
Criteo_f8	6.3(2.0)	9.7(1.3)	8.6(1.7)
Criteo_f8'	7.1(2.2)	9.5(1.4)	9.0(1.8)
Gerber_p2002	1.1(1.1)	1.0(1.2)	1.0(1.0)
Gerber_p2002'	0.6(1.1)	0.9(1.3)	0.7(1.2)
Gerber_p2004	0.8(1.1)	0.9(1.0)	0.7(1.3)
Gerber_p2004'	1.0(1.3)	0.9(1.3)	0.8(1.2)
Hillstrom_mens	5.5(2.6)	5.3(2.7)	5.5(2.7)
Hillstrom_mens'	6.5(2.7)	6.1(2.5)	6.2(1.9)
Hillstrom_newbie	6.2(2.7)	5.8(2.0)	6.3(2.4)
Hillstrom_newbie'	6.2(2.2)	6.2(2.4)	5.9(2.6)
Megafone_X16	3.5(0.4)	3.5(0.4)	3.6(0.4)
Megafone_X16'	3.6(0.4)	3.7(0.4)	3.5(0.4)
Megafone_X21	3.5(0.4)	3.4(0.4)	3.5(0.4)
Megafone_X21'	3.3(0.6)	3.2(0.6)	3.4(0.5)
Synth1	0.9(1.6)	1.2(0.7)	0.9(2.1)
Synth1'	2.1(0.7)	-0.3(0.8)	1.0(1.6)
Synth2	1.9(0.1)	1.9(0.1)	1.7(0.1)
Synth2'	1.7(0.0)	1.7(0.2)	1.9(0.1)
retailHero_age	1.2(0.3)	1.2(0.4)	1.2(0.3)
retailHero_age'	1.2(0.3)	1.2(0.3)	1.2(0.3)
retailHero_trNum	1.2(0.3)	1.2(0.4)	1.2(0.3)
retailHero_trNum'	1.2(0.4)	1.3(0.3)	1.2(0.3)
zenodoSynth_X10	12.6(1.9)	12.3(2.0)	12.1(1.8)
zenodoSynth_X10'	12.6(1.8)	12.3(1.5)	12.4(1.7)
zenodoSynth_X31	12.2(2.0)	12.5(1.7)	12.0(2.0)
zenodoSynth_X31'	12.2(1.8)	12.2(1.8)	12.6(1.8)

Slearner with Linear Regression

{}	SLearner_LinR	SLearner_LR_wt1	SLearner_LR_wt2
SyntheticData_zenodo_20Kprimex10_informative	0.1(2.2)	-0.1(2.1)	-0.1(2.1)
SyntheticData_zenodo_20Kprimex31_uplift_inc...	0.2(2.1)	0.1(2.3)	0.1(2.0)
SyntheticData_zenodo_20Kx10_informative	0.0(2.5)	-0.1(2.2)	-0.0(2.1)
SyntheticData_zenodo_20Kx31_uplift_increase	0.1(1.9)	0.1(2.1)	0.1(1.9)
criteoNewSample50kIdPf2	-0.2(1.9)	-0.1(2.2)	-0.1(2.6)
criteoNewSample50kIdPf8	-0.2(1.7)	-0.1(2.1)	-0.3(2.5)
criteoNewSample50kprimeIdPf2	-0.2(2.1)	-0.2(2.4)	-0.6(2.3)
criteoNewSample50kprimeIdPf8	-0.4(1.9)	-0.0(2.4)	-0.1(2.4)
gerber_Khiops_NvsCDIdPp2002	0.1(0.9)	0.2(1.0)	-0.1(1.0)
gerber_Khiops_NvsCDIdPp2004	0.0(1.1)	0.0(1.0)	0.0(1.0)
gerber_Khiops_NvsCDprimeIdPp2002	-0.2(0.9)	-0.0(0.9)	-0.1(1.0)
gerber_Khiops_NvsCDprimeIdPp2004	0.1(1.1)	-0.2(1.0)	-0.1(1.1)
hillstrom_w_Discmens	0.2(2.4)	-0.1(2.1)	0.1(2.2)
hillstrom_w_Discnewbie	-0.1(2.4)	0.3(2.3)	-0.4(2.3)
hillstrom_w_Discprimemens	0.0(2.2)	0.0(2.2)	0.1(2.1)
hillstrom_w_Discprimenewbie	0.2(2.4)	-0.2(2.5)	0.0(2.3)
megafoneIdPX_16	-0.0(0.6)	0.1(0.7)	0.0(0.5)
megafoneIdPX_21	-0.0(0.6)	0.0(0.5)	0.0(0.5)
megafoneprimeIdPX_16	-0.1(0.5)	0.1(0.5)	0.0(0.5)
megafoneprimeIdPX_21	0.0(0.5)	0.0(0.5)	0.0(0.5)
retailHeroPreProcessedMinaDiscage	-0.0(0.3)	-0.1(0.4)	-0.1(0.3)
retailHeroPreProcessedMinaDiscprimeage	-0.0(0.4)	0.0(0.3)	-0.0(0.3)
retailHeroPreProcessedMinaDiscprimetransactionN...	**0.0(0.3)	-0.0(0.4)	-0.0(0.4)
retailHeroPreProcessedMinaDisctransactionNumber	-0.0(0.4)	0.0(0.4)	0.0(0.3)
syntheticDataRandomMinaUnequalY_NOISEComb2	-0.1(0.4)	-0.0(0.4)	0.0(0.4)
syntheticDataRandomMinaUnequalY_NOISEprimeComb2	0.1(0.5)	-0.2(0.4)	0.0(0.3)
verysmallRateYNoisyv2Comb2	0.6(1.6)	0.4(1.3)	0.8(1.4)
verysmallRateYNoisyv2primeComb2	-1.0(1.6)	0.8(1.1)	0.5(1.8)

Slearner with Xgboost

{}	SLearner_Xgboost	SLearner_Xgboost_wt1	SLearner_Xgboost_wt2
SyntheticData_zenodo_20Kprimex10_informative	11.8(2.1)	11.4(2.4)	11.3(2.1)
SyntheticData_zenodo_20Kprimex31_uplift_inc...	11.1(2.1)	10.8(2.1)	11.0(2.4)
SyntheticData_zenodo_20Kx10_informative	11.5(2.0)	11.4(2.4)	11.1(2.2)
SyntheticData_zenodo_20Kx31_uplift_increase	11.7(2.2)	10.9(1.8)	11.1(2.1)
criteoNewSample50kIdPf2	8.0(1.9)	7.5(2.1)	8.6(2.5)
criteoNewSample50kIdPf8	8.4(1.9)	8.4(2.0)	8.4(2.1)
criteoNewSample50kprimeIdPf2	6.2(2.7)	6.8(1.8)	6.5(2.8)
criteoNewSample50kprimeIdPf8	6.9(2.1)	7.5(2.9)	7.3(2.0)
gerber_Khiops_NvsCDIdPp2002	-2.0(2.0)	-1.9(2.0)	-2.1(1.9)
gerber_Khiops_NvsCDIdPp2004	-1.8(2.0)	-1.7(2.1)	-2.2(1.8)
gerber_Khiops_NvsCDprimeIdPp2002	-1.9(2.2)	-2.1(1.9)	-1.7(1.7)
gerber_Khiops_NvsCDprimeIdPp2004	-1.9(2.0)	-1.8(2.0)	-1.8(2.3)
hillstrom_w_Discmens	2.5(2.7)	3.0(2.3)	3.1(2.5)
hillstrom_w_Discnewbie	3.3(2.2)	3.0(2.4)	2.3(2.0)
hillstrom_w_Discprimemens	2.9(2.9)	2.7(2.6)	2.3(2.4)
hillstrom_w_Discprimenewbie	2.5(2.7)	2.6(2.5)	3.7(2.4)
megafoneIdPX_16	17.9(0.4)	17.9(0.4)	17.8(0.5)
megafoneIdPX_21	18.4(0.4)	18.5(0.4)	18.2(0.5)
megafoneprimeIdPX_16	18.2(0.5)	18.2(0.5)	17.9(0.5)
megafoneprimeIdPX_21	18.3(0.4)	18.3(0.4)	18.0(0.5)
retailHeroPreProcessedMinaDiscage	0.9(0.3)	0.8(0.4)	0.7(0.4)
retailHeroPreProcessedMinaDiscprimeage	0.8(0.4)	0.8(0.4)	0.7(0.4)
retailHeroPreProcessedMinaDiscprimetransactionN...	0.6(0.4)	0.7(0.4)	0.7(0.4)
retailHeroPreProcessedMinaDisctransactionNumber	0.9(0.4)	0.9(0.3)	0.9(0.4)
syntheticDataRandomMinaUnequalY_NOISEComb2	10.1(0.1)	9.9(0.1)	9.8(0.1)
syntheticDataRandomMinaUnequalY_NOISEprimeComb2	9.9(0.2)	9.9(0.2)	10.0(0.2)
verysmallRateYNoisyv2Comb2	12.2(1.2)	11.8(1.4)	10.0(1.2)
verysmallRateYNoisyv2primeComb2	9.9(1.4)	10.8(1.3)	11.6(1.2)

Name		Name	Last commit message	Last commit date
Latest commit History 34 Commits
2MLR		2MLR
2MXgboost		2MXgboost
CT_LR_weighted/CTLR2		CT_LR_weighted/CTLR2
CT_Xgboost_weighted/CTXGboost2		CT_Xgboost_weighted/CTXGboost2
FigsBench/GIT_REPO		FigsBench/GIT_REPO
Synth		Synth
slearner_lr		slearner_lr
slearnerxgboost/Slearner_Xgboost		slearnerxgboost/Slearner_Xgboost
README.md		README.md
main.py		main.py

MinaWagdi/UpliftEvaluation_NRA

Folders and files

Latest commit

History

Repository files navigation