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/ UPCommons. Portal de...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/
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/
Recolector de Ciencia Abierta, RECOLECTA
Bachelor thesis . 2024
License: CC BY NC ND
versions View all 2 versions
addClaim

Análisis de tweets mediante modelos predictivos y procesamiento del lenguaje natural

Authors: Del Arco Llargués, Lucas;

Análisis de tweets mediante modelos predictivos y procesamiento del lenguaje natural

Abstract

Esta investigación se centra en el desarrollo de modelos de clasificación multiclase mediante técnicas de ciencia de datos aplicadas a un conjunto de tuits, asignándolos a 5 y 3 categorías según su nivel de positividad. Para llevar a cabo este proyecto de manera ordenada, se ha aplicado la metodología CRISP-DM. Se han desarrollado múltiples modelos y se han tomado diferentes métricas para asegurar que la distribución de sentimientos resultante sea lo más similar a la original y que el modelo tenga la mayor tasa de aciertos. Para esto, utilizaremos tres métricas: la divergencia KL, la F1 de cada una de las categorías y, por último, la .accuracy". Para construir estos modelos, hemos utilizado tres algoritmos de complejidad diferente: Naive Bayes, Redes Neuronales Transformer y BERT. En líneas generales, los modelos basados en algoritmos de mayor complejidad ofrecen una me- jor capacidad predictiva. Dependiendo de los objetivos de la empresa, un modelo más sencillo puede resultar suficiente y aportar prestaciones similares

Aquesta investigació es centra en el desenvolupament de models de classificació multiclasse mitjançant tècniques de la ciència de dades aplicades a un conjunt de tuits, assignant-los 5 i 3 categories segons el seu nivell de positivitat. Per dur a terme aquest projecte de manera ordenada, s’ha aplicat la metodologia CRISP-DM. S’han desenvolupat múltiples models i s’han pres diferents mètriques per assegurar que la dis- tribució de sentiments resultant sigui el més similar a l’original i que el model tingui la major taxa d’encerts. Per això, utilitzarem tres mètriques: la divergència KL, la F1 de cadascuna de les categories i, finalment, l’.accuracy". Per construir aquests models, hem utilitzat tres algorismes de complexitat diferent: Naive Bayes, Xarxes Neuronals Transformer i BERT. En general, els models basats en algorismes de major complexitat ofereixen una millor capacitat predictiva. Depenent dels objectius de l’empresa, un model més senzill pot ser suficient i aportar prestacions similars

This research focuses on the development of multiclass classification models using data scien- ce techniques applied to a set of tweets, assigning them to 5 and 3 categories based on their positivity level. To carry out this project in an organized manner, the CRISP-DM methodology has been applied. Multiple models have been developed, and different metrics have been taken to ensure that the resulting sentiment distribution is as similar to the original as possible, and that the model has the highest accuracy rate. For this purpose, three metrics will be used: KL divergence, F1 for each of the categories, and, finally, accuracy. To build these models, three algorithms of different complexity have been employed: Naive Bayes, Transformer Neural Networks, and BERT. Models based on more complex algorithms offer better predictive capabilities. Depending on the goals of the company, a simpler model may be sufficient and provide similar result

Country
Spain
Keywords

Xarxes socials en línia -- Informàtica, Online social networks -- Data processing, Àrees temàtiques de la UPC::Informàtica::Intel·ligència artificial::Aprenentatge automàtic, Automatic classification -- Software -- Design and construction, Classificació automàtica -- Programari -- Disseny i construcció, Anàlisi de contingut (Comunicació) -- Models matemàtics -- Disseny i construcció, Content analysis (Communication) -- Mathematical models -- Design and construction

  • 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
    OpenAIRE UsageCounts
    Usage byUsageCounts
    visibility views 84
    download downloads 208
  • 84
    views
    208
    downloads
    Powered byOpenAIRE UsageCounts
Powered by OpenAIRE graph
Found an issue? Give us feedback
visibility
download
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!
views
OpenAIRE UsageCountsViews provided by UsageCounts
downloads
OpenAIRE UsageCountsDownloads provided by UsageCounts
0
Average
Average
Average
84
208
Green