A Missing QoS Prediction Approach via Time-Aware Collaborative Filtering

Abstract

Quality of Service (QoS) guarantee is an important issue in building service-oriented applications. Generally, some <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> values of a service are unknown to its users who have never invoked the service before. Fortunately, collaborative filtering ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">CF</i> )-based methods are proved feasible for missing <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> prediction and have been widely used. However, these methods seldom took the temporal factors into consideration. Indeed, historical <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> values contain more information about user (or service) similarity. Furthermore, as the application environment is dynamic, obtained <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> values usually have short timeliness. Hence, using outdated <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> values will largely decrease the prediction accuracy. In order to resolve this issue, we proposed a time-aware collaborative filtering approach. First, we proposed a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> model to filter out outdated <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> values, and divided the obtained <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> values into several time slices. Then, we computed the average value of historical <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> as temporal <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> forecast. In addition, by introducing time-aware similarity computation mechanism, we succeeded to select real similar neighbor users (or services) and further predict the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">CF</i> -based <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> based on <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">CF</i> technology. Finally, we can predict the final missing <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> by combining temporal <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> forecast and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">CF</i> -based <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">QoS</i> prediction. Experiment results show that our approach can receive better prediction precision.