Open AccessHRCE: Detecting Food Security Events in Social Media
Author(s) -
Wang Gao,
Yuan Fang,
Yuwei Wang,
Fan Zhang
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1437/1/012090
Subject(s) - computer science , baseline (sea) , task (project management) , recall , social media , food security , recurrent neural network , artificial intelligence , deep learning , word (group theory) , event (particle physics) , f1 score , machine learning , artificial neural network , world wide web , psychology , cognitive psychology , geography , linguistics , oceanography , philosophy , physics , management , archaeology , quantum mechanics , economics , agriculture , geology
Analyzing food security events shared on social networks not only helps people deepen their understanding of food security events, but also helps managers cope with these events. In this paper, we propose a model that utilizes task-specific features and a deep learning model to detect food security events from tweets, called HRCE. Specifically, the proposed model leverages a hierarchical Recurrent Neural Network (RNN) with Long Short-Term Memory (LSTM) that takes word embeddings as inputs, and combines contextual embeddings to identify food security events from social media. We collected a novel food security related dataset from Twitter, and manually annotated 2,418 tweets. We conducted experiments on this dataset and concluded that HRCE outperforms baseline methods in terms of precision, recall and F1-score.