Behavioural Feature Extraction For Context-Aware Traffic Classification Of Mobile Applications
Traffic classification is becoming more complex due to proliferations of mobile applications coupled with growing diversity of traffic classes. This motivates the needs for improved traffic classification method that preserve classification accuracy while supporting more traffic classes. This the...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2018
|
| Subjects: | |
| Online Access: | http://eprints.usm.my/60904/1/Behavioural%20feature%20extraction%20for%20cut.pdf |
| Summary: | Traffic classification is becoming more complex due to proliferations of mobile
applications coupled with growing diversity of traffic classes. This motivates the needs
for improved traffic classification method that preserve classification accuracy while
supporting more traffic classes. This thesis identified domain-specific features that are
effective for accurate, large-scale and scalable mobile applications classification using
machine learning techniques. This thesis designed a context-aware traffic classification
framework that includes a set of sequential algorithms from cleaning datasets, to
identifying new features and detecting optimal classifier(s) based on problem contexts to
improve classification accuracy in multi-variate traffic classification. |
|---|