Personalized Ad Recommendation Systems for Life-Time Value Optimization with Guarantees

Authors: Georgios Theocharous, Philip S. Thomas, Mohammad Ghavamzadeh

IJCAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We apply these methods to a real PAR problem, both for evaluating the final performance and for optimizing the parameters of the RL algorithm. Our results show that a RL algorithm equipped with these offpolicy evaluation techniques outperforms the myopic approaches. For our experiments we used 2 data sets from the banking industry.
Researcher Affiliation Collaboration Georgios Theocharous Adobe Research theochar@adobe.com Philip S. Thomas UMass Amherst and Adobe Research phithoma@adobe.com Mohammad Ghavamzadeh Adobe Research and INRIA ghavamza@adobe.com
Pseudocode Yes Algorithm 1 GREEDYOPTIMIZATION(Xtrain, Xtest, δ, ϵ) : compute a greedy strategy using Xtrain, and predict the 1 δ lower bound on the test data Xtest and the value function.
Open Source Code No The paper does not provide an explicit statement or link indicating that the source code for the described methodology is publicly available.
Open Datasets No For our experiments we used 2 data sets from the banking industry. On the bank website when customers visit, they are shown one of a finite number of offers. The reward is 1 when a user clicks on the offer and 0, otherwise. We extracted/created features, in the categories shown in Table 1. The paper does not provide concrete access information for these datasets.
Dataset Splits Yes For both algorithms we start with three data sets an Xtrain, Xval and Xtest. We splitted the random strategy data into a test set and a validation set.
Hardware Specification No The paper does not provide specific details about the hardware (e.g., CPU, GPU models, memory) used for running the experiments.
Software Dependencies No The paper mentions software components and algorithms like random forest, FQI, Student's t-test, and BCa bootstrap, and even refers to 'MATLAB' but does not specify any version numbers for these software dependencies or libraries.
Experiment Setup Yes For all experiments we set γ = 0.9 and ϵ = 0.1.