Word Embeddings is a very powerful feature extraction technique for text data. In the case of Word Embeddings, every word is represented as a n-dimensional dense vector. And the beauty of Word Embeddings is in the way it manages to retain the semantic relationship among words, as every token has n-dimensional parameters explaining its characteristics.
Do check out our Movie Sentiment Analysis Project using Neural Networks, where we applied Word Embeddings for feature extraction of our IMDb Movie Reviews. Link: • Sentiment Analysis with LSTM | Deep Learni...