E-mail classification with machine learning and word embeddings for improved customer support

Abstract

Classifying e-mails into distinct labels can have a great impact on customer support. By using machine learning to label e-mails, the system can set up queues containing e-mails of a specific category. This enables support personnel to handle request quicker and more easily by selecting a queue that match their expertise. This study aims to improve a manually defined rule-based algorithm, currently implemented at a large telecom company, by using machine learning. The proposed model should have higher F 1-score and classification rate. Integrating or migrating from a manually defined rule-based model to a machine learning model should also reduce the administrative and maintenance work. It should also make the model more flexible. By using the frameworks, TensorFlow, Scikit-learn and Gensim, the authors conduct a number of experiments to test the performance of several common machine learning algorithms, text-representations, word embeddings to investigate how they work together. A long short-term memory network showed best classification performance with an F 1-score of 0.91. The authors conclude that long short-term memory networks outperform other non-sequential models such as support vector machines and AdaBoost when predicting labels for e-mails. Further, the study also presents a Web-based interface that were implemented around the LSTM network, which can classify e-mails into 33 different labels.

Exploring the effectiveness of word embedding based deep learning model for improving email classification

“…A DNN model to analyze emotive elements and other features to classify emails was proposed by Zamir et al (2020) and achieved an accuracy of 97.2%. The method to classify emails into separate labels and subsequently used to create queues holding emails belonging to a given category is proposed by Borg et al (2021). Continuous bag of words, Skipgram, N-gram and GloVe is used for feature extraction.…”

“…SVMs assume that the input data is linearly separable in geometric space, although this is not always the case. As a result, SVM maps the input to a high-dimensional feature space, where a linear decision boundary is built to maximize the difference between the two classes (Borg et al , 2021). Suppose a training set is not linearly separable.…”

A Text Classification Methodology to Assist a Large Technical Support System

“…Several approaches to help understand the human language to better respond to their demands have already been proposed in the NLP literature and achieved promising results [7], [9]- [11].…”

“…Borg et al [7] proposed the use of ML to classify emails in 33 categories to improve customer support in a telecommunication company. The authors affirmed that the e-mails' categorization could support the employees in selecting the e-mails that better match their expertise, resulting in quicker responses.…”

“…Large telecommunication companies have also been using NLP to classify e-mails into distinct labels to improve cus- VOLUME 4, 2016 tomer support [7]. This enables support personnel to handle requests quicker and more efficiently, establishing stable and positive relations with the customer.…”