Towards Practical Usage of a Domain Adaptation Algorithm in the Early Hours of a Disaster

Many machine learning techniques have been proposed to reduce the information overload in social media data during an emergency situation. Among such techniques, domain adaptation approaches present greater potential as compared to supervised algorithms because they don’t require labeled data from the current disaster for training. However, the use of domain adaptation approaches in practice is sporadic at best. One reason is that domain adaptation algorithms have parameters that need to be tuned using labeled data from the target disaster, which is presumably not available. To address this limitation, we perform a study on one domain adaptation approach with the goal of understanding how much source data is needed to obtain good performance in a practical situation, and what parameter values of the approach give overall good performance. The results of our study provide useful insights into the practical application of domain adaptation algorithms in real crisis situations.

[1]  Cornelia Caragea,et al.  Mapping moods: Geo-mapped sentiment analysis during hurricane sandy , 2014, ISCRAM.

[2]  Leysia Palen,et al.  Supporting “Everyday Analysts” in Safety- and Time-Critical Situations , 2011, Inf. Soc..

[3]  L. Palen,et al.  Crisis informatics—New data for extraordinary times , 2016, Science.

[4]  Muhammad Imran,et al.  Cross-Language Domain Adaptation for Classifying Crisis-Related Short Messages , 2016, ISCRAM.

[5]  Niloy Ganguly,et al.  Extracting Situational Information from Microblogs during Disaster Events: a Classification-Summarization Approach , 2015, CIKM.

[6]  Shady Elbassuoni,et al.  Practical extraction of disaster-relevant information from social media , 2013, WWW.

[7]  C. Castillo,et al.  Big Crisis Data: Social Media in Disasters and Time-Critical Situations , 2019 .

[8]  Starr Roxanne Hiltz,et al.  Red Tape: Attitudes and Issues Related to Use of Social Media by U.S. County-Level Emergency Managers , 2015, ISCRAM.

[9]  Leysia Palen,et al.  Natural Language Processing to the Rescue? Extracting "Situational Awareness" Tweets During Mass Emergency , 2011, ICWSM.

[10]  Nic Herndon,et al.  Empirical Study of Domain Adaptation Algorithms on the Task of Splice Site Prediction , 2014, BIOSTEC.

[11]  Hassan Sajjad,et al.  Rapid Classification of Crisis-Related Data on Social Networks using Convolutional Neural Networks , 2016, ICWSM 2016.

[12]  Leysia Palen,et al.  Online public communications by police & fire services during the 2012 Hurricane Sandy , 2014, CHI.

[13]  Sarah Vieweg,et al.  Processing Social Media Messages in Mass Emergency , 2014, ACM Comput. Surv..

[14]  David Yarowsky,et al.  Unsupervised Word Sense Disambiguation Rivaling Supervised Methods , 1995, ACL.

[15]  Kathleen M. Carley,et al.  Social Media in Disaster Relief , 2014 .

[16]  Cornelia Caragea,et al.  Twitter Mining for Disaster Response: A Domain Adaptation Approach , 2015, ISCRAM.

[17]  Cornelia Caragea,et al.  Disaster Response Aided by Tweet Classification with a Domain Adaptation Approach , 2018 .

[18]  Leysia Palen,et al.  Chatter on the red: what hazards threat reveals about the social life of microblogged information , 2010, CSCW '10.

[19]  Fernando Diaz,et al.  CrisisLex: A Lexicon for Collecting and Filtering Microblogged Communications in Crises , 2014, ICWSM.

[20]  Andrea H. Tapia,et al.  Good Enough is Good Enough: Overcoming Disaster Response Organizations’ Slow Social Media Data Adoption , 2014, Computer Supported Cooperative Work (CSCW).

[21]  Thomas Ludwig,et al.  Social Media and Emergency Services?: Interview Study on Current and Potential Use in 7 European Countries , 2015, Int. J. Inf. Syst. Crisis Response Manag..

[22]  Shanshan Zhang,et al.  Semi-supervised Discovery of Informative Tweets During the Emerging Disasters , 2016, ArXiv.