Open squaredcubed64 opened 1 year ago
We should download pre-trained word embeddings to improve our sentiment analysis. This lets us leverage language information from a massive text database, so our neural network can focus on analyzing tweets rather than understanding words.
We should also consider using word embeddings trained specifically on tweets because Twitter users use slang that rarely appears in English.
We should download pre-trained word embeddings to improve our sentiment analysis. This lets us leverage language information from a massive text database, so our neural network can focus on analyzing tweets rather than understanding words.