Predicting the Quality of Experience for Internet Video with Fuzzy Decision Tree

Abstract

In this paper, we attempt to predict users' quality of experience (QoE) with the log data collected from the web sites of Internet video service providers. To this end, we first collect service log data in the wild from one of the Top 5 most popular providers in China. Then we do a series of data preprocessing to format the original semi-structured log data to structured. We calculate several key video quality metrics, such as join time and frame rate, and explore the distributions of each quality metric, as well as the relationship between individual quality metric and user engagement. Considering that user engagement may be a result of comprehensive effect of several metrics, we apply fuzzy decision tree (FDT), a kind of classification algorithms in the area of machine learning, to develop the predictive model of user QoE for Internet video. Finally, we compare the prediction accuracy of our model with the model developed using decision tree on several different datasets. Our model separately achieves about 20% and 10% improvement in prediction accuracy on the dataset of sessions with the same content type and the dataset of sessions with mobile access devices.

Development of Fuzzy Relationships between Different Quality of Experience Parameters

“…The authors define zones in the -dimensional space where different values of various parameters lead to the same QoE rating. Zhang et al in [16] use fuzzy decision trees to predict user QoE from the log data collected from different Internet video service providers in China.…”

Modelling user quality of experience from objective and subjective data sets using fuzzy logic

“…The models and metrics presented here are all based on subjective test results obtained in controlled laboratory environments, with two exceptions: [9,50]. However, in [9], only short video clips were used for testing user perception, while in [50], the QoE is predicted from user engagement behavior and not from actual subjective quality tests.…”

“…However, in [9], only short video clips were used for testing user perception, while in [50], the QoE is predicted from user engagement behavior and not from actual subjective quality tests. This lack of assessment models that originate from uncontrolled experiments was our primary motivation in this study.…”

“…Based on the obtained results, the authors proposed an asymmetric adaptation model capable of mimicking human opinions when watching video with time-varying quality. Zhang et al in [50] use fuzzy decision trees to predict user QoE from the log data collected from different Internet video service providers in China. After processing the raw data sets (the sets contained information about video ID, content and video type, client IP address and location, access device, join time, frame rate, bandwidth, buffering times and buffering ratio), the authors model user engagement as a key aspect of viewer behavior that, they claim, in some sense reflects users' QoE.…”