This is the code, and the material needed to reproduced the result of our paper published at ECML 2018. Paper title: "A Comparison of Neural Network Methods for Accurate Sentiment Analysis of Stock Market Tweets"
result_matrix.xlsx contains the details of results from each neural network run along with baseline.
full_tokenized_DNN.csv contains the tokenized tweets used for preprocessing, tweet date, sentiment and ID.
For pre-processing, first run the script ./run_preprocessing Use train_model.py in python3 to run CNN or LSTM models. Pay attention to flags to change hyperparameters, or change LSTM flag.
This code, was built and modifed based on network available at https://github.com/bernhard2202/twitter-sentiment-analysis.