Dymanic Data Prefetching with ...

Mobile devices can provide the user location-dependent information services (LDISs), enhanced by prefetching the related information along with the route guidance based on user mobility mode. The current prefetching approaches are based on static assignments of probabilities of changing directions, hence the overall cache-miss for users are not reduced by the mobilities of the individuals. In this paper, we adopt the Prediction-by-Partial Match (PPM) technology to predict the best possible mobility of a user elaborated from his/her past history, to prefetch the most useful information for their LDIS. In our experiments, it is proven that our approach is more effective than Persone's approach for improving the rate of cache-miss and the energy consumption at most cases of prefetching data to get LDIS through on a route.