Efficient Nonparametric Subgraph Detection Using Tree Shaped Priors

Authors: Nannan Wu, Feng Chen, Jianxin Li, Baojian Zhou, Naren Ramakrishnan

AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, using extensive experiments we demonstrate the performance of our proposed algorithms in two real-world application domains (water pollution detection in water sensor networks and spatial event detection in social media networks) and contrast against state-of-the-art connected subgraph detection methods.
Researcher Affiliation Academia Dept. of Computer Science & Engineering, Beihang University, Beijing 100191, China Dept. of Informatics, University at Albany, SUNY, Albany, NY 12203 Dept. of Computer Science, Virginia Tech, Arlington, VA 22203
Pseudocode Yes Algorithm 1: Tree-Shaped-Priors Subgraph Detection
Open Source Code Yes Wu, N.; Chen, F.; Li, J.; Zhou, B.; and Ramakrishnan, N. Appendix, 2005: http://www.cs.albany.edu/ fchen/aaaiap.pdf.
Open Datasets Yes 1) Water Pollution Dataset. The Battle of the Water Sensor Networks (BWSN) (Ostfeld and Salomons 2008) provides a real-world network... 2) Event Detection Dataset. ...Gold Standard Reports (GSR) of 4279 official haze outbreak records (level 3) were collected from official websites (MEP ), and each GSR record was formatted as ( Time(YYYYMMDD) , Location(Province) )
Dataset Splits No The paper describes the datasets used and evaluation metrics, but it does not explicitly provide details about training, validation, or test dataset splits.
Hardware Specification No The paper does not provide specific details regarding the hardware used to run the experiments.
Software Dependencies No The paper mentions tools like EPANET and Weibo data, but it does not provide specific software dependencies with version numbers for reproducibility.
Experiment Setup Yes Specifically, for Event Tree and Graph-Laplacian, we tested the set of λ values: {0.1, 0.2, , 1.0, 50, 100, , 1500}. As Event Tree requires edge weights, we define the weight of an edge in the water pipeline network as the length of the pipeline segment to the edge and define the weight of an edge in the user-user network of Weibo as 1 without a better way. Two nonparametric scan statistics, HC and BJ, were evaluated. We set the parameters αmax 0.15 and the number of seed nodes C 5 for NPHGS and our methods.