Analyzing App Reviews: A ComparativeEvaluation of Machine Learning Algorithmson a Spanish Dataset

María Isabel Limaylla-Lunarejo ^[1] ; Nelly Condori-Fernandez ^[2] ; Miguel R. Luaces ^[1]
1. [1] Universidade da Coruña
  
  Universidade da Coruña
  
  A Coruña, España
2. [2] Universidade de Santiago de Compostela
  
  Universidade de Santiago de Compostela
  
  Santiago de Compostela, España
Localización: Proceedings XoveTIC 2024: Impulsando el talento científico / coord. por Manuel Lagos Rodríguez, Tirso Varela Rodeiro, Javier Pereira-Loureiro , Manuel Francisco González Penedo , 2024, págs. 67-72
Idioma: inglés
Enlaces
- Texto completo
Resumen
- Currently, the internet plays a main role in collecting and providing information on the needs and preferences of app users. App reviews contain valuable insights, such as bug reports, feature requests, and user feedback. However, manually analyzing these reviews is a time-consuming task. In this paper, we conducted an experiment to automate the process of analyzing app reviews using machine learning algorithms. We utilized and translated the dataset from Gu et al. (2015) to Spanish, which contains approximately 34,000 reviews from several apps. Three algorithms were trained: Multinomial Naive Bayes, Logistic Regression, and Support Vector Machine, with hyperparameter optimization performed via Grid Search. Logistic Regression achieved the highest performance with a maximum F1-score of 0.74.