A Mean-Variance Based Index for Dynamic Context Data Lookup

primary functionality of context aware applications is the retrieval of different types of context data from various context sources and adapting their behavior accordingly. In order to facilitate context aware application development, a context aware middleware must provide an effective context data management and lookup strategy. The use of a traditional index for indexing dynamic context data is not feasible due to the high update overhead. In this paper, we propose a context data indexing mechanism that utilizes the statistical properties of data viz. the mean and variance to cluster similar data values together and minimizes the need for frequent index updates. Experimental results indicate that the performance of the proposed index structure is satisfactory with respect to the query response time and query accuracy together with a low maintenance overhead.