Skip to content

Latest commit

 

History

History
54 lines (32 loc) · 2.5 KB

distant-reading.md

File metadata and controls

54 lines (32 loc) · 2.5 KB
Raw
description
And Similarity Analysis

Distant reading

[INTRO] In "Macroanalysis", Jackers suggests a new approach to computational analysis of texts, based on the concept of distant reading as defined by Franco Moretti.

This is no longer reading that we are talking about (Jockers 2013, 31)

The approach of macroanalysis takes inspiration from the difference between "microeconomic" and "macroeconomic", suggesting a quantifiable inspections of big corpora, that allows to capture the main phenomena and tendencies.

The underlying assumption is that by exploring the literary record writ large, we will better understand the context in which individual texts exist and thereby better understand those individual texts (Jockers 2013, 41)

Trying to analyse and contextualize "Entreprecariat", the distant reading is a difficult and controversial path, since there are no wide and reliable referrements to this new concept in the literature and this brings some difficulties in putting a corpus on the theme.

We decided to procede with a semantic analysis of the text, by means of word embeddings, in order to inspect quantified textual data.

[BASIC CONTEXT ANALYSIS] Our analysis started from a corpus that could contain the word 'entreprecariat', in which the author is Silvio Lorusso, out of the tweets.

We used the Python libraries spacy and gensim to process the text and do the word embeddings.

The pipeline we followed in order to pre-process all the texts is the following:

  • Preprocessing the texts with Spacy
  • Removing all the stopwords and punctuation
  • Sentence recognition
  • Word embedding with Word2Vec

[SIMILARITY ANALYSIS] For each word we assigned ten vectors, and against those vectorialization we made the following queries:

  • Which word is the most similar to entreprecariat?
  • Which is the most dissimilar?
from gensim.models import KeyedVectors

wv = KeyedVectors.load("vectors/word2vec.wordvectors", mmap='r')

#SIMILARITIES

print(wv.most_similar(""))

print("The similarity between ----- and ----- is:", wv.similarity("", ""))

#ANALOGY DIFFERENCE

#which word is to A as B is to C?
print(wv.most_similar(positive=["A", "B"], negative=["C"], topn=3))

It is possible to query the model on a Jupyter notebook provided on our repository, or you can download the folder from there:

{% file src=".gitbook/assets/InspectWordEmbedding.zip" %}