0

Text representation, in simple words, is representing text in sensible numeric form. You can read in detail from the following paragraph

Text representation is one of the fundamental problems in text mining and Information Retrieval (IR). It aims to numerically represent the unstructured text documents to make them mathematically computable. For a given set of text documents $D = \{d_i, i=1, 2,...,n\}$, where each $d_i$ stands for a document, the problem of text representation is to represent each $d_i$ of $D$ as a point $s_i$ in a numerical space $S$, where the distance/similarity between each pair of points in space $S$ is well defined.

But, I came across the phrase "text feature representation" in research papers. Features, in general, are present in dataset. But, I think, features can be characters or words or documents or complete text (as a single feature?) in the case of the text. I am not sure about what we call features in the text.

So, I am not sure about what is meant by text feature representation. Is it the same as text representation?

hanugm
  • 3,571
  • 3
  • 18
  • 50

1 Answers1

1

I think that literature is simply inconsistent in this regard. But here's a distinction that I think helps to shade a bit of light on this question:

text representation: as you said we have to convert text into numerical variables. This term refers to general strategies to convert text into numbers, like embedding, bag of words, and so on.

text features representation: this is what we feed to a real model. The difference is that it can be a combination of different text representation. For example, I could apply a TF-IDF vectorizer to a corpus and use it to encode each word of a sentence and then use also pre-trained embedding to encode the same sentence/document, concatenating for each word both vectors. Or another commonly used strategy is to use multiple n-grams, which also consist in applying multiple text representations to the same sentence/document.

Edoardo Guerriero
  • 5,153
  • 1
  • 11
  • 25