|
|
Bimonthly Since 1986 |
ISSN 1004-9037
|
|
|
|
|
Publication Details |
Edited by: Editorial Board of Journal of Data Acquisition and Processing
P.O. Box 2704, Beijing 100190, P.R. China
Sponsored by: Institute of Computing Technology, CAS & China Computer Federation
Undertaken by: Institute of Computing Technology, CAS
Published by: SCIENCE PRESS, BEIJING, CHINA
Distributed by:
China: All Local Post Offices
|
|
|
|
|
|
|
|
|
|
Abstract
In an IT service management environment, manual classification of support tickets may involve misclassification and hence results in assigning the ticket to a wrong resolver group. There is a need to develop an automated ticket classifier system which does the auto categorization of service desk tickets. This research work proposes a methodology to develop such an automated ticket classifier by exploring the unsupervised machine learning techniques. The proposed method uses the text document clustering approach to categorize the IT service desk tickets by mining the natural language ticket description of the unlabelled tickets submitted by the end users. Non-Negative Matrix Factorization based document clustering algorithm is used to generate the ticket clusters. The cluster generation is then followed by labelling of clusters by extracting the most frequent terms in each cluster. Each ticket cluster represents a single class label and the generated ticket cluster model is further used to categorize the new unlabelled service desk ticket. A real-world IT infrastructure unlabelled service desk ticket dataset is used for the experimental purposes. The cluster performance evaluation metrics like Number of clusters, Entropy and Davies-Bouldin Index are used to evaluate the proposed ticket cluster model.
Keyword
IT Service desk, Ticket classification, Unsupervised machine learning, Non-Negative Matrix Factorization, Davies-Boudin index.
PDF Download (click here)
|
|
|
|
|