Identifying emergency stages in facebook posts of police departments with convolutional and recurrent neural networks and support vector machines

Abstract

Classification of social media posts in emergency response is an important practical problem: accurate classification can help automate processing of such messages and help other responders and the public react to emergencies in a timely fashion. This research focused on classifying Facebook messages of US police departments. Randomly selected 5,000 messages were used to train classifiers that distinguished between four categories of messages: emergency preparedness, response and recovery, as well as general engagement messages. Features were represented with bag-of-words and word2vec, and models were constructed using support vector machines (SVMs) and convolutional (CNNs) and recurrent neural networks (RNNs). The best performing classifier was an RNN with a custom-trained word2vec model to represent features, which achieved the F1 measure of 0.839.

Deep Learning for Multi-Class Identification From Domestic Violence Online Posts

“…In the context of crisis management and response during natural disasters (earthquake [62], and flood [6]), CNNs were adopted to classify the social media posts as either informative or non-informative, and resulted in improved performance over the traditional classifiers such as SVM, LR and RF. In the other study regarding online posts classification in the emergency situations, RNNs outperformed the CNNs and SVM [63]. For various NLP applications such as sentiment analysis and questionanswering [64], GRUs and LSTMs proved superior over the CNNs.…”

Deep Learning for Multi-Class Antisocial Behavior Identification From Twitter

“…CNNs demonstrated significantly improved performance in these scenarios compared to Random Forest, Linear Regression and Support Vector Machine. In a similar emergency post-classification study, RNN's outperformed Support Vector Machine and Convolutional Neural networks in [38], LSTM outperformed CNN [14], and GRUs outperformed both CNN and RNNs in its vanilla form [39]. For many Natural language processing applications such as question-answering and sentiment analysis [40], LSTMs and GRUs demonstrated better performance over CNNs [41].…”

Didn’t roger that: Social media message complexity and situational awareness of emergency responders

“…RNN demonstrated the best performance among several classifiers on this dataset, including convolutional neural networks and support vector machines (SVM). Further details on classifier selection for this dataset are described in Pogrebnyakov and Maldonado (2017).…”