Complaint management is a problem faced by many organizations that is both vital to customer image and highly dependent on human resources. This work attempts to tackle a part of the problem, by classifying summaries of complaints using machine learning models in order to better redirect these to the appropriate responders. The main challenges of this task is that training datasets are often small and highly imbalanced. This can can have a big impact on the performance of classification models. The dataset analyzed in this work suffers from both of these problems, being relatively small and having labels in different proportions. In this work, two different techniques are analyzed: combining classes together to increase the number of elements of the new class; and, providing new artificial examples for some classes via translation into other languages. The classification models explored were the following: k-NN, SVM, Naïve Bayes, boosting, and Deep Learning approaches, including transformers. The paper concludes that although, as expected, the classes with little representation are hard to classify, the techniques explored helped to boost the performance, especially in the classes with a low number of elements. SVM and BERT-based models outperformed their peers.

This work was partly funded through national funds by FCT - Fundação para a Ciência e Tecnologia, I.P. under project UIDB/04466/2020 (ISTAR).

Text classification,Natural language processing,Deep learning,BERT

• Ciências da Computação e da Informação - Ciências Naturais
• Engenharia Eletrotécnica, Eletrónica e Informática - Engenharia e Tecnologia