The first stage to ameliorate and optimize the provision of mobile streaming services is calculating and estimating the Quality of Experience (QoE) of a multimedia stream. Machine learning models offer a solution to obtain the complicated relationships between various influencing factors and QoE. This article presents a novel online QoE prediction algorithm, namely, Stacked Incremental Support Vector Machine model (SISVM). This model relies on a combination of two main procedures: Ensemble learning model, more specifically, Stacking model and Incremental Support Vector Machine (ISVM). The principal contribution of this work consists of the use of separate Multiclass Incremental SVM, in order to estimate different parts of the problem and to combine the outputs of each model using probabilistic techniques. In fact, using ISVM in real-time applications is best fit for the manipulation non-stationary data. Experimental results have proven clearly that the proposed model has outperformed the existing state-of-the-art models, in terms of classification accuracy and computational complexity.
[1]
Robert Tibshirani,et al.
Classification by Pairwise Coupling
,
1997,
NIPS.
[2]
Abdelhamid Mellouk,et al.
Survey on machine learning-based QoE-QoS correlation models
,
2014,
2014 International Conference on Computing, Management and Telecommunications (ComManTel).
[3]
Abdul Hameed,et al.
A Decision-Tree-Based Perceptual Video Quality Prediction Model and Its Application in FEC for Wireless Multimedia Communications
,
2016,
IEEE Transactions on Multimedia.
[4]
Klaus-Robert Müller,et al.
Incremental Support Vector Learning: Analysis, Implementation and Applications
,
2006,
J. Mach. Learn. Res..