On Identifying Disaster-Related Tweets: Matching-Based or Learning-Based?

To, Hien; Agrawal, Shruti; Kim, Seon Ho; Shahabi, Cyrus

doi:10.1109/bigmm.2017.82

Cited by 47 publications

(29 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, we compare our results with the learning-based algorithm employed by To et al [45], who also evaluated their model's performance with CrisisLexT26 datasets. In particular, their learning-based approach used Word2Vec, TF-IDF, latent semantic indexing, and logis- Table 3).…”

Section: Discussionmentioning

confidence: 98%

“…Several classification approaches have been developed to identify relevant and irrelevant social media information, such as clustering [5,6], keyword matching [45], and term-vector similarity [12]. However, to the best of our knowledge, no existing work in this area includes interactive learning with real-time data, focusing instead on improving the machine learning algorithms themselves [5, 16, 23, 31-33, 38, 45, 46, 54] or interactively training on archived datasets [9,18].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Interactive Learning for Identifying Relevant Tweets to Support Real-time Situational Awareness

Snyder

Lin

Karimzadeh

et al. 2019

IEEE Trans. Visual. Comput. Graphics

View full text Add to dashboard Cite

InterfaceAutomated Label Verify Label Fig. 1. Our interactive learning framework allows users to train text relevance classifiers in real-time to improve situational awareness.In this example, a real-time tweet regarding a car accident is incorrectly classified as "Irrelevant". Through the SMART 2.0 interface, the user can view its label and correct it to "Relevant", thereby retraining and improving the classifier for incoming streaming data.Abstract-Various domain users are increasingly leveraging real-time social media data to gain rapid situational awareness. However, due to the high noise in the deluge of data, effectively determining semantically relevant information can be difficult, further complicated by the changing definition of relevancy by each end user for different events. The majority of existing methods for short text relevance classification fail to incorporate users' knowledge into the classification process. Existing methods that incorporate interactive user feedback focus on historical datasets. Therefore, classifiers cannot be interactively retrained for specific events or user-dependent needs in real-time. This limits real-time situational awareness, as streaming data that is incorrectly classified cannot be corrected immediately, permitting the possibility for important incoming data to be incorrectly classified as well. We present a novel interactive learning framework to improve the classification process in which the user iteratively corrects the relevancy of tweets in real-time to train the classification model on-the-fly for immediate predictive improvements. We computationally evaluate our classification model adapted to learn at interactive rates. Our results show that our approach outperforms state-of-the-art machine learning models. In addition, we integrate our framework with the extended Social Media Analytics and Reporting Toolkit (SMART) 2.0 system, allowing the use of our interactive learning framework within a visual analytics system tailored for real-time situational awareness. To demonstrate our framework's effectiveness, we provide domain expert feedback from first responders who used the extended SMART 2.0 system.Index Terms-Interactive machine learning, human-computer interaction, social media analytics, emergency/disaster management, situational awareness

show abstract

Section: Discussionmentioning

confidence: 98%

Section: Introductionmentioning

confidence: 99%

Interactive Learning for Identifying Relevant Tweets to Support Real-time Situational Awareness

Snyder

Lin

Karimzadeh

et al. 2019

IEEE Trans. Visual. Comput. Graphics

View full text Add to dashboard Cite

show abstract

“…However, the existing literature on disaster analysis lacks in dealing with the authenticity of the content, and additional measures need to be taken to check the authenticity of news and other information shared in social media. Moreover, the domain also lacks of public datasets, e.g., for Twitter most of the works [110,124,80,125] use self-collected datasets. Moreover, the datasets are often not large enough in terms of total number of images and types of the natural disaster events they cover.…”

Section: Open/key Research Challengesmentioning

confidence: 99%

“…In order to extract the location of the disaster from the tweets, the authors use a geo-tagged filter, while K-Nearest Neighbour (K-NN) algorithm has been adopted for the disaster distribution analysis. The approach proposed in [124] relies on the use of matching keywords and hashtags for the identification of relevant messages from the social media streams. Moreover, the authors also provide a comparative analysis of their proposed matching keywords approach against a learning-based system.…”

Section: Disaster Detection In Twitter Textmentioning

confidence: 99%

Natural disasters detection in social media and satellite imagery: a survey

Said

Ahmad

Riegler

et al. 2019

Multimed Tools Appl

108

View full text Add to dashboard Cite

The analysis of natural disaster-related multimedia content got great attention in recent years. Being one of the most important sources of information, social media have been crawled over the years to collect and analyze disasterrelated multimedia content. Satellite imagery has also been widely explored for disasters analysis. In this paper, we survey the existing literature on disaster detection and analysis of the retrieved information from social media and satellites. Literature on disaster detection and analysis of related multimedia content on the basis of the nature of the content can be categorized into three groups, namely (i) disaster detection in text; (ii) analysis of disaster-related visual content from social media; and (iii) disaster detection in satellite imagery. We extensively review different approaches proposed in these three domains. Furthermore, we also review benchmarking datasets available for the evaluation of disaster detection frameworks. Moreover, we provide a detailed discussion on the insights obtained from the literature review, and identify future trends and challenges, which will provide an important starting point for the researchers in the field.

show abstract

“…The automatic classification of disaster-related tweets is a relevant and complex problem, still open, addressed by many researchers (e.g., [31][32][33][34][35][36]). About the difficulties to be overcome, Stowe, K. et al ( [33], p. 1) wrote: "Identifying relevant information in social media is challenging due to the low signal-to-noise ratio."…”

Section: The App Templatementioning

confidence: 99%

A Citizen-Sensing-Based Digital Service for the Analysis of On-Site Post-Earthquake Messages

Felice

Iessi

2019

IJGI

View full text Add to dashboard Cite

The effectiveness of disaster response depends on the correctness and timeliness of data regarding the location and the impact of the event. These two issues are critical when the data come from citizens’ tweets, since the automatic classification of disaster-related tweets suffers from many shortcomings. In this paper, we explore an approach based on participatory sensing (i.e., a subset of mobile crowdsourcing that emphasizes the active and intentional participation of citizens to collect data from the place where they live or work). We operate with the hypothesis of a “friendly world”, that is by assuming that after a calamitous event, in the survivors prevails the feeling of helping those who suffer. The extraction, from the Twitter repository, of the few tweets relevant to the event of interest has a long processing time. With the aggravating circumstance in the phase that follows a severe earthquake, the elaboration of tweets clashes with the need to act promptly. Our proposal allows a huge reduction of the processing time. This goal is reached by introducing a service and a mobile app, the latter is an intermediate tool between Twitter and the citizens, suitable to assist them to write structured messages that act as surrogates of tweets. The article describes the architecture of the software service and the steps involved in the retrieval, from the Twitter server, of the messages coming from citizens living in the places hit by the earthquake; moreover, it details the storage of those messages into a geographical database and their processing using SQL.

show abstract

On Identifying Disaster-Related Tweets: Matching-Based or Learning-Based?

Cited by 47 publications

References 23 publications

Interactive Learning for Identifying Relevant Tweets to Support Real-time Situational Awareness

Interactive Learning for Identifying Relevant Tweets to Support Real-time Situational Awareness

Natural disasters detection in social media and satellite imagery: a survey

A Citizen-Sensing-Based Digital Service for the Analysis of On-Site Post-Earthquake Messages

Contact Info

Product

Resources

About