This project aims to perform sentiment analysis on text data using a Recurrent Neural Network (RNN) implemented with Keras. Sentiment analysis involves determining the sentiment expressed in a piece of text, such as positive, negative, or neutral.
The model is a Sequential model, which contains a linear stack of neural layers. The first layer is a simple RNN layer which consists of 50 neurons, whereas the second layer is a Dense layer with 6 neurons. Finally, the last layer is a softmax activation layer. After the neural network is ready, KerasCLassifier is used as the wrapper to help train the model with the training data.
To use the model to determine sentiment of a text, the text string will first have to be tokenized and then padded. After storing that as a list, you can use the model to predict the sentiment