Automated categorization of support tickets using machine learning
(2018) In LU-CS-EX 2018-14 EDAM05 20181Department of Computer Science
- Abstract
- Customer Support is usually an expensive service for a lot of companies.
At Telavox, a majority of the support tickets are sent via e-mail. A support
ticket is sent by a customer, containing a request for support. Support tickets
from customers could reveal valuable information. Additional insights can
be given by labelling the tickets. Today, employees at the customer support
department at Telavox assign labels to the tickets manually. In this thesis, we
explore whether it is possible to automatically classify support tickets from a
predetermined set of labels based on their textual representation. We explore
different data representations, classification algorithms, clustering algorithms
and neural nets to find the best... (More) - Customer Support is usually an expensive service for a lot of companies.
At Telavox, a majority of the support tickets are sent via e-mail. A support
ticket is sent by a customer, containing a request for support. Support tickets
from customers could reveal valuable information. Additional insights can
be given by labelling the tickets. Today, employees at the customer support
department at Telavox assign labels to the tickets manually. In this thesis, we
explore whether it is possible to automatically classify support tickets from a
predetermined set of labels based on their textual representation. We explore
different data representations, classification algorithms, clustering algorithms
and neural nets to find the best solution. We found that a ridge regression
classifier gave the best result. We also carried out a qualitative analysis of the
corpus and found that it is not always easy to classify tickets by looking only
at the text. Finally, we develop an application that could be used to predict
what label is best suited for a given support ticket. Using the probability, our
model can rank the predictions. The correct label is presented in the three first
predicted labels 91.6% of the time. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/8962734
- author
- Andréason, Hanna LU and Nilsson, Christopher
- supervisor
- organization
- course
- EDAM05 20181
- year
- 2018
- type
- H2 - Master's Degree (Two Years)
- subject
- keywords
- Classification, Clustering, Scikit-learn, Text Categorization, support tickets
- publication/series
- LU-CS-EX 2018-14
- report number
- LU-CS-EX 2018-14
- ISSN
- 1650-2884
- language
- English
- id
- 8962734
- date added to LUP
- 2018-12-13 15:35:02
- date last changed
- 2018-12-13 15:35:02
@misc{8962734, abstract = {{Customer Support is usually an expensive service for a lot of companies. At Telavox, a majority of the support tickets are sent via e-mail. A support ticket is sent by a customer, containing a request for support. Support tickets from customers could reveal valuable information. Additional insights can be given by labelling the tickets. Today, employees at the customer support department at Telavox assign labels to the tickets manually. In this thesis, we explore whether it is possible to automatically classify support tickets from a predetermined set of labels based on their textual representation. We explore different data representations, classification algorithms, clustering algorithms and neural nets to find the best solution. We found that a ridge regression classifier gave the best result. We also carried out a qualitative analysis of the corpus and found that it is not always easy to classify tickets by looking only at the text. Finally, we develop an application that could be used to predict what label is best suited for a given support ticket. Using the probability, our model can rank the predictions. The correct label is presented in the three first predicted labels 91.6% of the time.}}, author = {{Andréason, Hanna and Nilsson, Christopher}}, issn = {{1650-2884}}, language = {{eng}}, note = {{Student Paper}}, series = {{LU-CS-EX 2018-14}}, title = {{Automated categorization of support tickets using machine learning}}, year = {{2018}}, }