Thesis Análisis del sentir SANSANO en tiempos de pandemia mediante técnicas de machine learning y visualización de datos
Loading...
Date
2023-10
Authors
Journal Title
Journal ISSN
Volume Title
Program
Ingeniería Civil Informática
Departament
Campus
Campus Santiago San Joaquín
Abstract
Este trabajo de memoria se centra en el análisis de sentimientos de las confesiones de estudiantes de la Universidad Técnica Santa María publicadas en la red social Instagram durante pandemia. Para lograrlo, se aplicaron técnicas de aprendizaje automático, específicamente utilizando modelos de procesamiento del lenguaje natural, algoritmos de clúster y visualizaciones de datos. Los resultados destacan que las emociones y sentimientos expresados en los textos tienden a ser más negativos que positivos. Además, se identificó que el algoritmo Fuzzy-C Means con un parámetro m=1.01 se desempeñó como el mejor modelo de clusterización entre los probados. La relevancia de este trabajo radica en su potencial para comprender las necesidades de los alumnos y ofrecer apoyo temprano a aquellos que puedan requerirlo. En un contexto de creciente importancia de la salud mental, la detección anticipada de posibles signos de angustia o depresión a través del análisis de texto en las publicaciones de Instagram se convierte en una herramienta valiosa.
This thesis work focuses on sentiment analysis of confessions shared by students of Universidad Técnica Santa María on the platform Instagram during pandemic. To achieve this, machine learning techniques were applied, specifically utilizing natural language processing models, clustering algorithms, and data visualizations. The results reveal that the emotions and sentiments expressed in the texts tend to be more negative than positive. Additionally, it was identified that the Fuzzy-C Means algorithm with a parameter m=1.01 performed as the best clustering model among those tested. The significance of this work lies in its potential to comprehend the needs of students and provide early support to those who may require it. In a context of increasing emphasis on mental health, the early detection of possible signs of distress or depression through text analysis in Instagram posts becomes a valuable tool.
This thesis work focuses on sentiment analysis of confessions shared by students of Universidad Técnica Santa María on the platform Instagram during pandemic. To achieve this, machine learning techniques were applied, specifically utilizing natural language processing models, clustering algorithms, and data visualizations. The results reveal that the emotions and sentiments expressed in the texts tend to be more negative than positive. Additionally, it was identified that the Fuzzy-C Means algorithm with a parameter m=1.01 performed as the best clustering model among those tested. The significance of this work lies in its potential to comprehend the needs of students and provide early support to those who may require it. In a context of increasing emphasis on mental health, the early detection of possible signs of distress or depression through text analysis in Instagram posts becomes a valuable tool.
Description
Keywords
Procesamiento de lenguaje natural, Procesamiento de datos, Algoritmos computacionales
