Header menu link for other important links
Embedded Bi-directional GRU and LSTMLearning Models to Predict Disasterson Twitter Data
, J.T. Jones Thomas, P. Kesavan
Published in Elsevier B.V.
Volume: 165
Pages: 511 - 516
The deep learning techniques namely Long Short Term Memory (LSTM) network and Bi-directional Gated Recurrent Unit (BGRU) network turn to be de facto to build an optimal assembly line for neural network models. The prevailing state-of-the-art approaches require a substantial amount of labeled data detailed to an unambiguous event in the training phase. In this paper, embedded bi-directional GRU and LSTM learning models is applied for disaster event prediction that uses deep learning techniques to categorize the tweets. The performance of the proposed neural network model is evaluated on CrisisLexT26 benchmarking dataset. The resulting validation accuracy is estimated by comparing LSTM and bi-directional GRU with and without word embeddings. The experiments demonstrate the model selector choose the deep learning techniques to predict the disaster event with reasonably high accuracy. © 2019 Procedia Computer Science. All rights reserved.
About the journal
JournalData powered by TypesetProcedia Computer Science
PublisherData powered by TypesetElsevier B.V.