Tracking Disaster Footprints with Social Streaming Data

Authors: Lu Cheng, Jundong Li, K. Selcuk Candan, Huan Liu370-377

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experimental results on real-world datasets collected during Hurricane Harvey and Florence reveal the effectiveness of our framework. In this section, we conduct qualitative and quantitative analyses to evaluate the performance of TDF for finding common and distinct topics during disaster response.
Researcher Affiliation Academia 1Computer Science and Engineering, Arizona State University, USA 2Department of Electrical and Computer Engineering, University of Virginia, USA 3Department of Computer Science & School of Data Science, University of Virginia, USA
Pseudocode Yes Algorithm 1 The proposed TDF framework.
Open Source Code No Datasets and select pieces of custom code are available upon request.
Open Datasets No We crawled real-world datasets related to two recent natural disasters Hurricane Harvey (2017) and Hurricane Florence (2018) from Twitter... Data and select pieces of custom code are available upon request.
Dataset Splits No The paper mentions a batch size for incoming data but does not explicitly provide train/validation/test dataset splits for reproducibility.
Hardware Specification No The paper does not explicitly describe the hardware used to run its experiments.
Software Dependencies No The paper mentions using TF-IDF, NMF, and the Tweet Tracker system, but does not specify version numbers for any software dependencies or libraries used in the experiments.
Experiment Setup Yes Values of k and kc are set to 10 and 7 respectively, kd = k kc = 3. ... The best performance is achieved when α lies between [500,1000] and β is between [0.1,1]. ... All the quantitative results are computed with kc = 7, α = 1000, β = 0.1.