Incorporating Implicit Link Preference Into Overlapping Community Detection

Authors: Hongyi Zhang, Irwin King, Michael Lyu

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

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate our PNMF model on several real-world networks. Experimental results show that our model outperforms state-of-the-art approaches and can be applied to large datasets.
Researcher Affiliation Academia 1Shenzhen Key Laboratory of Rich Media Big Data Analytics and Applications, Shenzhen Research Institute, The Chinese University of Hong Kong, Shenzhen, China 2Department of Computer Science and Engineering, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong
Pseudocode Yes Algorithm 1 Community detection using PNMF
Open Source Code No The paper does not provide any statements about releasing its source code or a link to a code repository for the methodology described.
Open Datasets Yes We examine our model with several benchmark datasets available on the Internet. ... For the first category, we choose nine undirected networks collected by Newman1 as our datasets. For the second category, three large datasets from SNAP2 are used. ... 1http://www-personal.umich.edu/mejn/netdata 2http://http://snap.stanford.edu/data/
Dataset Splits Yes We first reserve 10% of links as validation set. Then we vary p and learn model parameters with the remaining 90% of links for each p. After that, we use the node-community membership matrix F to generate the adjacency matrix G and use G to predict the links in validation set according to our motivation that linked pairs have higher value than non-linked pairs in G. Finally, we pick the p with the best prediction score as our pre-assigned number of communities.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., GPU models, CPU types, memory) used to run the experiments.
Software Dependencies No The paper mentions implementing the model using stochastic gradient descent but does not specify any software names with version numbers (e.g., Python, PyTorch, TensorFlow versions, or other libraries).
Experiment Setup No The paper describes how certain parameters (number of communities 'p' and membership threshold 'β') are chosen, mentioning a range for β '[0.7, 0.8]' and the use of cross-validation. It also mentions 'learning rate and the regularization parameter are appropriate' but does not provide specific numerical values for these hyperparameters or other training configurations.