I have done some reading. I want to implement an LSTM with pre-trained word embeddings (I also have plans to create my word embeddings, but let's cross that bridge when we come to it).
In any given sentence, you don't usually need to have all the words as most of them do not contribute to the sentiment, such as the stop words and noise. So, let's say there is a sentence. I remove the stop words and anything else that I deem unnecessary for the project. Then I run the remaining words through the word embedding algorithm to get the word vectors.
Then what? How does it represent the sequence or the sentence 'cause it's just vector for a word.
For example, take the sentence:
The burger does not taste good.
I could remove certain words and still retain the same sentiment like so:
Burger not good.
Let's assume some arbitrary vectors for those three words:
Burger: $[0.45, -0.78, .., 1.2]$
not: $[9.6, 4.0, .., 5.6]$
good: $[3.5, 0.51, 0.8]$
So, those vectors represent the individual words. How do I make a sentence out of them? Just concatenate them?