Models of automatic recognition of collaborative emotions in rural secondary education institutions in Puno, 2024

Abstract

Nowadays, collaborative emotions expressed in social networks by teenagers are topics of educational and communication interest fundamental in human interactions, whose analysis has experienced an exponential growth in the digital era. The aim of the article is to determine a machine learning and/or deep learning model that performs better in the recognition of complex sequential text patterns of emotions expressed on social networking platforms by secondary school students in rural areas of Puno. The methodology comprises keyword sensing technique (KST) and natural language processing (NLP), and has been applied to a total of 2160 emotions expressed in comments on daily educational activities among friends and classmates on personal Facebook profile. Pre-processing, training, evaluation and implementation have been carried out in Google Colab with pandas, Scikit-learn, numpy, TensorFlow/Keras, Matplotlib and Seaborn libraries. As a result, the implementations of the ANN, SVM, Naive Bayes, XGBoost, Random Forest, KNN and LightGBM algorithms were compared and evaluated with the metrics of precision, recall, F1-score and AUC. ANN was found to be the best fitting model for the dataset, with an AUC of 97% correctly classifying emotions such as ‘Surprise’ and ‘Anger’, but has limitations with ‘Happiness’ and ‘Sadness’, emotions that are often confused with each other, especially with ‘Distress’. In conclusion, ‘Surprise’ highlights feelings expressed in the personal Facebook profile of students from rural Puno, where the Peruvian educational model does not meet the expectations of the highland youth who still preserve their own living culture of the Andean cosmovision, allowing them to survive in situations of poverty, a culture that instills in them the importance of solidarity and reciprocity, while ‘Anger’ reflects permanent frustrations due to marginalisation in basic sanitation services, motivating them to undertake migratory adventures to the cities of the coast. Once there, they suffer racial mistreatment, physical or psychological harassment and, emotionally, they lose self-esteem, unable to integrate into a full society with quality of life.