Prediction Model for User's QoE in Imbalanced Dataset

Abstract

Nowadays, the users of IPTV require better user experience and the media providers are interested in finding the key factors which will influence the Quality of Experience (QoE) and the way to predict the QoE. In this paper, we discuss the relationship between the status of IPTV set-top box and user's QoE. We first clean the dataset and conduct some statistical analysis in dataset. Then, we select the important features in key performance indicator (KPI) dataset by the feature selection technology. In addition, we compare the decision tree model and Adaboost model in KPI -- Cuser's complaint dataset which is a classical imbalanced dataset. Extensive experimental results demonstrate that in imbalanced KPI-user's complaint dataset Adaboost model performs better than the decision tree in terms of predicting the user's complaint, and importantly we design a cost coefficient for Adaboost model to achieve a higher accuracy.