SA-unsupervised-data

Sentiment Analysis of Books and Its impact on children using Deep Learning

Download the datasets and do follow [inal_lstm.ipynb and final_naivebayes.ipynb]

1. three kinds of data is here [actual data - all the specific words in the book] [data - raw data is in a list form in a data frame] [Final data - it is labeled with sentiments manually]

2. LSTM - Accruacy was 82.50%

3. Naive Bayes - Accuracy was 80% {18% negativity, 77.3% neutral and 4.7% positive }