TS-SS similarity for Answer Retrieval from Document in Python
Introduction
This article focuses on answer retrieval from a document by using a similarity algorithm. This task falls under Natural Language Processing which is a subset of Deep Learning. In this article, we will be understanding why do we require better techniques and what are the drawbacks of using naive algorithms. Moreover, we will be implementing a similarity-based technique for answer retrieval from the document.
This article is a follow-up to NLP: Answer Retrieval from Document using Python. I strongly recommend giving it a fast read.
Why do we need a new matrix?
Problems with Euclidean Distance
If two data vectors have no component values in common, they may have a smaller distance than the other pair of data vectors containing the same component values.
Let’s say for example we have three vectors → A (1, 0 , 0) , B (0 , -2 , -2) , C (10 , 0