Row-clustering of a Point Process-valued Matrix

Authors: Lihao Yin, Ganggang Xu, Huiyan Sang, Yongtao Guan

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The effectiveness of the proposed framework is demonstrated through simulation studies and real data analyses.
Researcher Affiliation Academia Institute of Statistics and Big Data, Renmin University, Beijing, China, 2Department of Statistics, Texas A&M University, College Station, Texas, 3University of Miami, Coral Gables, Florida
Pseudocode Yes Algorithm 1 Learning of the Single-level model in (2)
Open Source Code Yes Our code can be accessed via https://github.com/Lihao Yin/MMMPP.
Open Datasets Yes The City of Chicago collected the information of all taxi rides in Chicago since 2013 1. Each trip record in the dataset consists of drivers encrypted IDs, pick-up/drop-off time stamps, and locations (in the form of latitude/longitude coordinates). Footnote 1: https://data.cityofchicago.org/Transportation/Taxi-Trips/wrvz-psew
Dataset Splits Yes We evaluate and compare clustering stability based on a measure called clustering consistency via K-trial cross-validations [23, 24]... We compare the performance of DF, DMHP, and MM-MPP in terms of clustering consistency for three data sets with K = 100 trials.
Hardware Specification Yes Table 2: Running Time (in seconds) on Synthetic Data lists: GPU-ES (RTX 8000 48G GPU), CPU-ES (i7-7700HQ CPU), CPU-EM (i7-7700HQ CPU).
Software Dependencies No The paper mentions software like the 'rtweet' API [12] and 'NumPy' [7] but does not provide specific version numbers for any software dependencies used in the experiments.
Experiment Setup Yes Section 5.1 'Setting' states: 'We set the number of clusters C from 2 to 5 and set the number of accounts in each cluster to 500. We experiment with an increasing number of replicates (m = 1, 20 or 100)'. Additionally, it notes: 'For each iteration, 10,000 MCMC samples are drawn to approximate (10)'.