Powered by OpenAIRE graph
Found an issue? Give us feedback
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/ LAReferencia - Red F...arrow_drop_down
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
versions View all 1 versions
addClaim

Análisis de contenido de texto basado en procesamiento de lenguaje natural con BERT

Authors: González, Jairo; Angulo, Jesús; Andrés, Meza;

Análisis de contenido de texto basado en procesamiento de lenguaje natural con BERT

Abstract

Este artículo se centra en tratar de aliviar los problemas relacionados con el análisis de contenido. Discutiremos el uso de diferentes modelos de clasificación en el aprendizaje automático. Adoptamos este enfoque para resolver algunos problemas relacionados con el análisis cualitativo, como la fiabilidad en el tiempo y la disminución de la mano de obra cualificada. Lo hacemos para automatizar un proceso que suele requerir cantidades considerables de tiempo y recursos, como humanos capacitados y largos plazos de entrega. Exploramos el uso de diferentes técnicas como Random Forest y K-Nearest Neighbor, también probamos diferentes métodos de bolsa de palabras para codificar el texto. También evaluamos un prototipo de la solución propuesta con Representaciones de Codificación Bidireccional de Transformadores (BERT) bajo un conjunto de datos para la detección de noticias falsas debido a las limitaciones de alcance, sin embargo, es aplicable a otro corpus y otro contexto de texto. Finalmente, con los servicios de AWS implementaremos un sistema para la creación de una API que pueda ser utilizada por el usuario común e implementada en sus sistemas de clasificación.

This article focuses on trying to alleviate problems related to content analysis. We will discuss the use of different models for classification in Machine learning. We take this approach to solve some problems related to qualitative analysis, such as reliability over time and the decline of skilled labor. We do this to automate a process that usually requires considerable amounts of time and resources, such as trained humans and long lead times. We explored the use of different techniques like Random Forest and K-Nearest Neighbor, we also tried different bag of words methods to encode the text. We also evaluated a prototype of the proposed solution with Bidirectional Encoding Representations of Transformers (BERT) under a dataset for detection of fake news due to scope limitations, However, it is applicable to another corpus and other text context. Finally, with AWS services we will implement a system for the creation of an API that can be used by the common user and implemented in their classification systems.

Country
Colombia
Related Organizations
Keywords

text classification, clasificación de texto, PLN, NLP, BERT

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    0
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Average
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
0
Average
Average
Average
Green