How to handle new words that are in Test data but are not feature vectors in Word Embedding Matrix?

Installations & Python Packages
1

Course Name: Natural Language Processing in Trading, Section No: 9, Unit No: 4, Unit type: Notebook

 

2

Hi Gaurav,



The Out-Of-Vocabulary (OOV) words are a limitation of the word embedding techniques. The alternative to that can be using a lexicon-based approach like VADER, or maybe even combine embedding techniques with lexicon-based approaches.



Also, you may look at this post on how to handle this problem.



Hope this helps!