REVISTA DYNA NEW TECHNOLOGIES

Key words:

Polaridad sentimental, Análisis de sentimientos, Validación cruzada, Conjuntos de entrenamiento y prueba, Sentiment polarity, Sentiment analysis, Cross-validation, Training and test sets.

Article type:

ARTICULO DE INVESTIGACION / RESEARCH ARTICLE

Section:

RESEARCH ARTICLES

ABSTRACT
Today, social networking services have experienced exponential growth and have become an integral part of users' daily lives. These platforms, such as Twitter, have acquired a relevant role in the generation and dissemination of information in different segments of the population. The value of the information generated on these platforms has increased significantly in line with this increase in their use.
In this paper, we present a study that focuses on analyzing the polarity of a dataset extracted from Twitter. The goal is to understand which preprocessing techniques and classification methods can help us classify the polarity of messages in these unstructured datasets. To perform automatic identification of misogynistic sentiments on Twitter, experiments are conducted using different learning methods, such as Support Vector Machine, Naive Bayes, Logistic Regression, KNN and Random Forest. These methods are applied in two classification scenarios: cross-validation and training and test sets.
The results obtained demonstrate the feasibility of the proposed methodology and contribute to a theoretical-practical study to identify misogynistic messages in unstructured texts present in social networking platforms. This research seeks to provide a deeper understanding of the context of this current issue, taking into account the use of machine learning systems to identify the polarity of a text based on the emotions expressed by its author.
Keywords: Sentiment polarity, Sentiment analysis, Cross-validation, Training and test sets.