Telling Peer Direct Effects from Indirect Effects in Observational Network Data

Authors: Xiaojing Du, Jiuyong Li, Debo Cheng, Lin Liu, Wentao Gao, Xiongren Chen, Ziqi Xu

ICML 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on two semi-synthetic datasets derived from real-world networks and on a dataset from a recommendation system confirm the effectiveness of our approach.
Researcher Affiliation Academia 1University of South Australia, Adelaide, Australia 2School of Computing Technologies, RMIT University, Melbourne, Australia. Correspondence to: Jiuyong Li <EMAIL>, Debo Cheng <EMAIL>.
Pseudocode Yes D. The g DIS Algorithm. The pseudo-code of g DIS for estimating PDE, PIE and STE is presented in Algorithm 1.
Open Source Code No The paper does not provide explicit access to source code for the methodology described. It mentions using "Tensor Flow (Abadi et al., 2016)" and "Network X (Hagberg et al., 2008)" but these are third-party libraries, not the authors' own code release.
Open Datasets Yes We applied g DIS to two semi-synthetic real-world datasets, Blog Catalog (BC) and Flickr (Li et al., 2015), to evaluate its effectiveness on estimating PDE, PIE and STE. ... Blog Catalog2 (BC): ... 2https://www.blogcatalog.com/ ... Flickr3: ... 3https://www.flickr.com/ ... we conducted experiments using the Coat Dataset1, which simulates MNAR (missing not at random) data from online coat purchases. As described in (Schnabel et al., 2016), the dataset includes user features and 5-point ratings ... 1https://www.cs.cornell.edu/ schnabts/ mnar/
Dataset Splits Yes The resulting partitioned data distribution is shown in Table 5. Table 5. Data distribution after graph partitioning. Each tuple (m, k) indicates the number of nodes (m) and the feature dimension (k). Dataset Train Validation Test Blog Catalog (1722, 10) (1733, 10) (1731, 10) Flickr (1557, 10) (2526, 10) (1829, 10)
Hardware Specification No The paper does not provide specific hardware details used for running its experiments. It mentions software libraries but no CPU, GPU, or memory specifications.
Software Dependencies No The paper mentions that the model is implemented using "Python libraries Tensor Flow (Abadi et al., 2016) and Network X (Hagberg et al., 2008)", but it does not specify the version numbers for these software components.
Experiment Setup Yes In this section, we provide additional details about the experimental setup and configurations. Our proposed model is implemented using the Python libraries Tensor Flow (Abadi et al., 2016) and Network X (Hagberg et al., 2008). We performed a grid search on the validation set to select the key parameters (learning rate, hidden dimensions, and regularization strength) that yielded the best performance, as shown in Table 6. Table 7 summarizes the parameter settings used for peer effect analysis on Blog Catalog and Flickr.