KTH_bigdata_project

Twitter sentiment analysis

Naive Bayes

Simply run bayes_classifiers.py

SVM

Run svm_testing.py

LSTM

dir : KTH_bigdata_project/sentiment_analysis_tensorflow-master/

for train, run: python train.py

for predict, run: python predict.py --checkpoints_dir
for example: python predict.py --checkpoints_dir checkpoints/1481294288

for more information: see the readme file inside the project

CNN

1. Run Train_CNN.py for training process

2. Checkpoints will be saved in runs folder

3. Run Test_CNN_v2_final.py for test process

Hints: You should change the directory of input data in the test and train process based on your own data directory
text_cnn.py contain the model for CNN. data_helpers.py and data_input_v3.py doing some process related to input data of CNN