| || || Mobile computing -- Data processing|
| || || Big data analysis for mobile computing research project-semantic analysis for location awareness|
Author: Singh, Ashika Vandhana
Institution: University of the South Pacific.
Subject: Context-aware computing, Mobile computing -- Data processing, Global system for mobile communications -- Data processing
Call No.: Pac QA 76 .5915 .S56 2015
Copyright:Under 10% of this thesis may be copied without the authors written permission
Abstract: The mobile computing research phenomenon implicates optimized performance and enhanced smart information accessibility focusing on ease of anywhere, anytime information access using smart devices. Context awareness capability of smart devices is an insatiable requirement of smart device users, providing users with location awareness capabilities, using satellite communication to either broadcast their location, look for places of interest to them or finding their way using global positioning systems. The main focus of this research work is to explore the capabilities of different machine learning algorithms in building a context-aware model. The contextawareness is based on location parameters of a mobile device. The context-aware model is evaluated using the “citywide” dataset— that is provided by CRAWDAD (an online resource community for research) for benchmarking. The dataset contains real, long-term data collected from three participants using a Place Lab Client, which collects location coordinates of the participants as they move around a specified area. The traces collected are from laptops and Personal Digital Assistants (PDAs). In this experiment we explore the K-Means and K-Medoids algorithms to generate the clusters for each context. The stratified sampling techniques are used to generate the test sets. The test dataset is used to create data models for the semantic analysis. The results obtained in the experiment indicate that it is plausible to add semantics to mobile computing data for locations awareness. The results reveal that these machine learning algorithms are potentially the candidate solutions to identify places of significance to the user of a mobile device. These algorithms can be used to build context-aware model for the mobile-devices when a context is represented by the location of the device.