Sarcasm is tough to understand, so much more when the person in front is good with his/her words. Some sentences and even words spoken in certain contexts are sarcastic while they read as perfectly fine as a standalone. This project aims to enhance sarcasm detection by hybridizing the three best-performing models in past research work: CNN, LSTM, and Random Forest along with performing dynamic Learning Rate Schedule Optimization in order to cancel out noise in the data and help the deep learning models learn more effectively.
Clone this repository via
git clone https://github.com/mansheelagarwal/Contextual-Sarcasm_Detection.git