Page 241 - Big data - Concept and application for telecommunications

Big data - Concept and application for telecommunications 5

6.7 Big-data-driven quality of experience/quality of service improvement

User quality of experience (QoE) can be promoted by analysing big data from network status data and user
patterns.
Typically, various services and applications are managed by using a set of QoS parameters (e.g., packet loss,
delay, and jitter). However, management can be more efficient when the quality as perceived by end users
(i.e., QoE) is taken as the optimization objective instead of QoS. Towards this end, automatic and accurate
estimation of QoE in real time is the first step. Data analytics can help with QoE modelling and monitoring in
a diverse heterogeneous environment, which is essential for global network optimization.
As shown in Figure 6-6, the data needed to estimate QoE come from both the network and users. Besides
the technical factors, various non-technical factors exist that may influence QoE results, including: device
type; user emotion; habit; and expectation. Thus, in QoE evaluation, it is useful to create an individual profile
for each user, which is a user model representing user preferences, habits and interests. A user does not
usually like to spend much time answering questions to create a profile model. As an alternative, a user
profile can be built and monitored using data analytics with implicit information gathered by a profile
collection engine. The activities of users are tracked and compared to identify similarities and differences.
For example, the output of the motion detector in the profile collection engine may include (but is not limited
to) the number of clicks and scrolling on the screen. In the emotion detector, a user emotion may be
extracted from a detected user behaviour with affective computing techniques. Meanwhile, network data
including QoS parameters are collected through the measurement and signalling in the network. All data are
stored in a database for further processing. A machine-learning engine is then used to establish the
relationship between the influencing factors and the QoE through artificial intelligence. Machine-learning
techniques enable ever more accurate decision-making over time, even when the data sets are incomplete
or new situations arise. The analysis of large data sets leads to insights into users' real experiences, which
may need to incorporate social data. Data analytics can discover what operators need to know, which impacts
QoE across devices, services and network resources.

Figure 6-6 – Big-data-driven quality of experience improvement

Network and infrastructure 233