Method of Assessment and Improvement the Quality of Multimedia Services

摘要

Paper presents an improved method for assessing the quality of multimedia services through the use of a hierarchical neuro-fuzzy network. It is shown that artificial intelligence methods allows the hybrid system to take into account users' opinion and extract knowledge from the input stream of service quality indicators. Process of the comprehensive QoS assessment is presented. It combines analysis of objective measurements of the network operation characteristics and the data obtained as a result of user surveys. Work is based on the ITU-T E.802 recommendations, where it is proposed to use a four markets model to assess the quality of multimedia services. Proposed method takes into account the ratio of the increase in the quality indicator value and the cost of its correction. Importance, or "weight" of each of the quality indicators and quality components that shows their impact on the final result is also considered. Article presents an example of the quality indicators level correction for the service "streaming video". Based on MATLAB, a GUI application was developed that implements the work of the developed quality of services assessment method.