Thesis Análisis de sentimientos con técnicas de machine learning en la red social x respecto a instituciones públicas chilenas: Carabineros y Congreso Nacional
No Thumbnail Available
Date
2024
Authors
Journal Title
Journal ISSN
Volume Title
Program
Ingeniería Civil Informática
Departament
Campus
Campus Santiago San Joaquín
Abstract
El presente trabajo de memoria trata del análisis de sentimientos realizado sobre las opiniones de chilenos en X (Twitter) sobre las instituciones de Carabineros y el Congreso Nacional, en los periodos de tiempo que albergan el estallido social y el posterior primer plebiscito constitucional. Este análisis de sentimientos se ha basado en técnicas de aprendizaje automático, las cuales usan modelos para el procesamiento del lenguaje natural. Los resultados de este análisis y este trabajo resaltan la disminución de la opinión negativa de un 81 % a un 58 % y el aumento de la opinión neutra y positiva, en menor medida, en el caso de Carabineros. En el caso del Congreso Nacional, los resultados indican que los sentimientos de las opiniones se han mantenido en esos 3 años, con valores entre el 73 %, el 20 % y 6 % para los sentimientos negativos, neutros y positivos, respectivamente. Este trabajo de investigación pretende encontrar las razones de estos resultados y el porqué estas opiniones han cambiado de esta manera a lo largo de esos 3 años.
This work deals with the sentiment analysis of the opinions by Chileans on X (Twitter) about the institutions of Carabineros and the National Congress, in the time periods that host the social outburst and the subsequent first constitutional plebiscite. This sentiment analysis has been based on machine learning techniques, which use models for natural language processing. The results of this analysis and this work highlight the decrease of negative opinion from 81 % to 58 % and the increase of neutral and positive opinion, to a lesser extent, in the case of Carabineros. In the case of the National Congress, the results indicate that the sentiments of opinions have been maintained in those 3 years, with values between 73 %, 20 % and 6 % for negative, neutral and positive sentiments, respectively. This research work aims to find the reasons for these results and why these opinions have changed in this way over these 3 years.
This work deals with the sentiment analysis of the opinions by Chileans on X (Twitter) about the institutions of Carabineros and the National Congress, in the time periods that host the social outburst and the subsequent first constitutional plebiscite. This sentiment analysis has been based on machine learning techniques, which use models for natural language processing. The results of this analysis and this work highlight the decrease of negative opinion from 81 % to 58 % and the increase of neutral and positive opinion, to a lesser extent, in the case of Carabineros. In the case of the National Congress, the results indicate that the sentiments of opinions have been maintained in those 3 years, with values between 73 %, 20 % and 6 % for negative, neutral and positive sentiments, respectively. This research work aims to find the reasons for these results and why these opinions have changed in this way over these 3 years.
Description
Keywords
Carabineros, Análisis de sentimientos, Congreso Nacional, Procesamiento del lenguaje natural
