Why need to convert words/sentences to vectors in NLP

Suitable as an input for machine and deep learning

Photo by Boitumelo Phetla on Unsplash

It is a very important question in the field of natural language processing. One of the prime reasons is the use of linear algebra. The other reason is that most of the architecture in machine learning and deep learning uses input in the form of some values rather than a string.

There should be a way to convert the words into vectors.

For example:

I am Amit

This sentence is not suitable as an input for machine and deep learning. Someone will think that we give input as a sentence in deep learning algorithms but one thing to be noticed is that in the middle of the processes in the deep learning models they convert to the vectors only before feeding to the model.

The basic use case in natural language processing is sentiment analysis. Suppose the function takes input in a form of sentence/comment then it will predict the sentiment of a sentence/comment. The machine learning will apply the non-linear function to the comment and predict the sentiment. The input should be in a format of vectors for machine learning input.

If we see the sentiments of the comments in visual forms then we need to convert the positive comments…