Satirical News Detection and Analysis

Code for the paper

Satirical News Detection and Analysis using Attention Mechanism and Linguisitc Features

Dataset

Currently, the dataset is hosted at Mendeley.

Please download the data from Here.

The code requires linguitsitc features to be pre-processed.

We will provide the scripts for pre-processing and update dependencies soon.

Theano 0.9 Lasagne 0.2 python 3.4

char_hierarchical_linguistic_rnn.py - 4 level char-word-paragraph-document network with paragraph-level and document-level linguistic features.
char_hierarchical_para.py - 4 level char-word-paragraph-document network with paragraph-level linguistic features
char_hierarchical_doc.py - 4 level char-word-paragraph-document network with document-level linguistic features
char_hierarchical.py - 4 level char-word-paragraph-document network
hierarchical_linguistic_rnn.py - 3 level word-paragraph-document network with linguistic features
hierarchical_rnn.py - Implementation of Hierarchical Attention Network
arg_parser.py
data_utilities.py
evaluations.py
minibatch.py
networks.py
theano_function.py
utilities.py

Please see arg_parser.py for detail.

For more question about the dataset, please contact Fan Yang, fyang11@uh.edu