AUTOMATED HELP DESK SYSTEM BASED ON DEEP LEARNING

Year 2022, , 318 - 327, 21.12.2022

Metin Yılmaz , Efnan Şora Günal

https://doi.org/10.31796/ogummf.1038486

Abstract

A help desk is an organization's point of contact that provides a centralized information and support management service to its employees or customers. For the efficiency of the organization, it is of great importance that the queries coming to the help desk are grouped into the correct categories and directed to the right people on time. Therefore, in this study, an automatic help desk system based on deep learning is proposed. The proposed system automatically categorizes queries according to the sentences in their titles. Word embedding method was used for this process. After the text preprocessing steps, learning is performed in three layers (embedding, flatten, and dense) and the category to which the help desk queries belong is determined. For this purpose, IT help desk queries belonging to a corporate company were used. The dataset, consisting of a total of 28.104 requests in nine different categories, is divided into 60% training, 20% validation, and 20% test set. As a result of the experiments, the classification accuracy reaching 98% revealed that the proposed model is a good candidate for an automated help desk system.

Keywords

Deep Learning, Natural Language Processing, Text Classification, Help Desk

DERİN ÖĞRENME TEMELLİ OTOMATİK YARDIM MASASI SİSTEMİ

Abstract

Yardım masası, bir organizasyonun çalışanlarına veya müşterilerine merkezi bilgi ve destek yönetimi hizmeti sağlayan iletişim noktasıdır. Organizasyonun verimliliği açısından, yardım masasına gelen taleplerin doğru kategorilere ayrılarak, doğru kişilere ve zamanında yönlendirilmesi büyük önem arz etmektedir. Bu sebeple, bu çalışma kapsamında, derin öğrenmeye dayalı otomatik bir yardım sistemi önerilmiştir. Önerilen sistem, talepleri, başlıklarında yer alan cümlelere göre otomatik olarak uygun kategorilere ayırmaktadır. Bu işlem için kelime gömme (ing. word embedding) yöntemi kullanılmıştır. Metin ön işleme adımlarından sonra, üç katmanda (embedding, flatten ve dense) öğrenme gerçekleştirilerek, yardım masası taleplerinin ait olduğu kategori belirlenmektedir. Bu amaçla, kurumsal bir şirkete ait BT yardım masası talepleri kullanılmıştır. Dokuz farklı kategoride toplam 28.104 talepten oluşan veri kümesi, %60 eğitim, %20 doğrulama ve %20 test kümesine ayrılmıştır. Yapılan deneyler sonucunda %98’e ulaşan sınıflandırma doğruluğu, önerilen modelin otomatik bir yardım masası sistemi için iyi bir aday olduğunu ortaya koymuştur.

Keywords

Derin Öğrenme, Doğal Dil İşleme, Metin Sınıflandırma, Yardım Masası, Deep Learning Natural Language Processing Text Classification Help Desk, Deep Learning, Natural Language Processing, Text Classification, Help Desk

References

• ALRashdi, R., & O'Keefe, S. (2019). Deep learning and word embeddings for tweet classification for crisis response. arXiv preprint arXiv:1903.11024.
• Bian, J., Gao, B., & Liu, T. Y. (2014). Knowledge-powered deep learning for word embedding. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases, 132-148.
• Borko, H., & Bernick, M. (1963). Automatic document classification. Journal of the ACM (JACM), 10(2), 151-162.
• Cai, S., Palazoglu, A., Zhang, L., & Hu, J. (2019). Process alarm prediction using deep learning and word embedding methods. ISA Transactions, 85, 274-283.
• Habibi, M., Weber, L., Neves, M., Wiegandt, D. L., & Leser, U. (2017). Deep learning with word embeddings improves biomedical named entity recognition. Bioinformatics, 33(14), i37-i48.
• Jason Brownlee, (2017) How to Use Word Embedding Layers for Deep Learning with Keras. Eişim Adresi: http://machinelearningmastery.com/use-word-embedding-layers-deep-learning-keras.
• Jurafsky, Daniel; H. James, Martin (2000). Speech and language processing: an introduction to natural language processing, computational linguistics, and speech recognition. Upper Saddle River, N.J.: Prentice Hall. ISBN 978-0-13-095069-7.
• Keras, (2021), Sequential_model. Erişim adresi: http://keras.io/guides/sequential_model.
• Kilimci, Z. H., & Akyokus, S. (2018). Deep learning-and word embedding-based heterogeneous classifier ensembles for text classification. Complexity.
• Kocmi, T., & Bojar, O. (2017). An exploration of word embedding initialization in deep-learning tasks. arXiv preprint arXiv:1711.09160.
• Kowsari, K., Brown, D. E., Heidarysafa, M., Meimandi, K. J., Gerber, M. S., & Barnes, L. E. (2017). Hdltex: Hierarchical deep learning for text classification. In 16th IEEE International Conference on Machine Learning and Applications (ICMLA), 364-371.
• Li, S., Hu, J., Cui, Y., & Hu, J. (2018). DeepPatent: patent classification with convolutional neural networks and word embedding. Scientometrics, 117(2), 721-744.
• Loper, E., & Bird, S. (2002). Nltk: The natural language toolkit. arXiv preprint cs/0205028.
• Mikolov, T., Chen, K., Corrado, G., & Dean, J. (2013). Efficient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781.
• Powers, D. M. (2020). Evaluation: from precision, recall and F-measure to ROC, informedness, markedness and correlation. arXiv preprint arXiv:2010.16061.
• Semberecki, P., & Maciejewski, H. (2017). Deep learning methods for subject text classification of articles. In IEEE Federated Conference on Computer Science and Information Systems (FedCSIS), 357-360.
• Wang, J. H., Liu, T. W., Luo, X., & Wang, L. (2018, October). An LSTM approach to short text sentiment classification with word embeddings. In 30th Conference on Computational Linguistics and Speech Processing (ROCLING 2018), 214-223.
• Zhang, Y., Lu, J., Liu, F., Liu, Q., Porter, A., Chen, H., & Zhang, G. (2018). Does deep learning help topic extraction? A kernel k-means clustering method with word embedding. Journal of Informetrics, 12(4), 1099-1117.