Mobility

A Prediction-Based Scheme for Wireless Push Systems, Using a Statistical Hidden Markov Model

Download Now Free registration required

Executive Summary

In this paper a Hidden Markov Model (HMM) is applied at the broadcast server in order to provide accurate decisions for a wireless push system environment with unknown client demands. Clients are organized into groups and may request a set of items. The server side sends an item per time and then clients respond with a feedback, if the transmitted item is the desirable one. The novel model tries to adapt quickly and precisely to the dynamic changes of the clients' demands. Initially, the suggested scheme learns the intentions of the connected clients for a specific number of broadcasts and then begins to predict the clients' requests. Finally, the server side follows the results of the prediction procedure, by sending the appropriate items.

  • Format: PDF
  • Size: 169.3 KB