A Hybrid Machine Learning Model for Efficient Classification of IT Support Tickets in The Presence of Class Overlap.

Classifying customer support tickets according to the desired criteria is an important task in IT service management. Accurate classification enables the support agents to reuse similar previous resolutions which. in turn, reduces the ticket resolution time and enhances customer satisfaction. However, for large-scale IT corpora with hundreds of classes organized in a hierarchy, the task of accurate classification of classes at the higher level in the hierarchies is crucial to avoid errors propagating to the lower levels. One of the biggest challenges is the presence of a large number of shared words between different classes. This problem is widely known as overlapping classes. Misclassification due to overlapping regions is a critical problem that is not well addressed in the NLP field. In this paper, we detect overlapping classes from an ML algorithm perspective and propose a hybrid machine learning model based on a linear SVM classifier and a set of N hand-crafted rules to classify the incoming ticket with high accuracy where N is the number of overlapped classes. The experimental results on four datasets show that the proposed hybrid model achieves major improvements in terms of the F­score of the overlapped classes. Hence, we recommend that for text classification tasks with overlapping classes, linear SVM along with a set of handcrafted rules can provide an interpretable and superior performance for the misclassified classes.