Additional Multi-Touch Attribution for Online Advertising

Authors: Wendi Ji, Xiaoling Wang

AAAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on a large real-world advertising dataset illustrate that the our proposed method is superior to state-of-the-art techniques in conversion rate prediction and the credit allocation based on AMTA is reasonable.
Researcher Affiliation Academia Shanghai Key Laboratory of Trustworthy Computing, East China Normal University 3663 North Zhongshan Road, Shanghai, China wendyg8886@gmail.com, xlwang@sei.ecnu.edu.cn
Pseudocode No The paper provides mathematical equations and descriptions of methods but does not include any clearly labeled pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any concrete access information (e.g., a specific link or explicit statement) for the source code of the methodology described.
Open Datasets No The dataset is described as a 'real-world competition dataset provided by Miaozhen, a leading marketing technique company in China.' However, no direct URL, DOI, specific repository name, or citation to a public resource for accessing this dataset is provided.
Dataset Splits Yes the results are generated by 4-fold cross-validation over the users.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., CPU, GPU models, memory) used to run the experiments, only mentioning the use of 'mini-batch stochastic gradient descent' for optimization.
Software Dependencies No The paper mentions implementing 'a gradient descent algorithm' and using 'mini-batch stochastic gradient descent', but it does not specify any software names with version numbers.
Experiment Setup Yes The optimization method is taken on the regularized negative log likelihood with respect to parameters of p, λ, and Λ: arg min Θ L (Θ) + μ; In our experiments, we have used mini-batch stochastic gradient descent to reduce the communication cost; We predict whether a user will convert in a specified upcoming period (30, 15 and 7 days); the results are generated by 4-fold cross-validation over the users; we sampled 1% negative users for model training.