Fast Estimation of Causal Interactions using Wold Processes
Authors: Flavio Figueiredo, Guilherme Resende Borges, Pedro O.S. Vaz de Melo, Renato Assunção
NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our approach, called GRANGER-BUSCA, is validated on nine datasets. This is an advance in relation to most prior efforts which focus mostly on subsets of the Memetracker data. Regarding accuracy, GRANGER-BUSCA is three times more accurate (in Precision@10) than the state of the art for the commonly explored subsets Memetracker. and 5 Results and Experiments and We evaluate GRANGER-BUSCA and the aforementioned three baselines on 9 different datasets |
| Researcher Affiliation | Academia | Flavio Figueiredo Guilherme Borges Pedro O. S. Vaz de Melo Renato Assunc ao Universidade Federal de Minas Gerais (UFMG) {flaviovdf, guilherme.borges, olmo, assuncao}@dcc.ufmg.br |
| Pseudocode | Yes | In Algorithm 1 we show how Ogata s Modified Thinning algorithm [38] is adapted for GRANGERBUSCA. |
| Open Source Code | Yes | Reproducibility: http://github.com/flaviovdf/granger-busca |
| Open Datasets | Yes | Datasets: We evaluate GRANGER-BUSCA and the aforementioned three baselines on 9 different datasets, all of which were gathered from the Snap Network Repository5. ... 5https://snap.stanford.edu/data/ and Memetracker data [29] |
| Dataset Splits | No | The paper does not explicitly provide details on train/validation/test dataset splits, such as percentages, sample counts, or specific predefined splits for reproducibility. |
| Hardware Specification | Yes | Experiments were performed on a dedicated Azure Standard DS15 v2 instance with 20 Intel Xeon CPU E5-2673 v3 cores and 140GB of memory. |
| Software Dependencies | Yes | The code for each method is publicly available via the library tick4. ... 4https://github.com/X-Data Initiative/tick. Results produced with version 0.4.0.0. |
| Experiment Setup | Yes | Training is performed until convergence or until 300 iterations is reached. Our MCMC sampler executes for 300 iterations with αp = 1/K and β = 1. and For Hawkes-Granger, we fit the model with M = 10 basis functions as suggested in [1]. |