User-Centric Contexts Sensing for Instant Information Seeking Support

Use of portable devices receives a tremendous growth in recent years. Such device becomes a universal portal, which connects users and the Web together, in addition to its ordinary built purposes. Users tend to share things via available channels (and/or social media as well) with the convenience brought by the device. This phenomenon, however, brings to the publics an open issue about the management of a vast amount of data, or user generated content as its specific term, and its potential usage. So this research, from a perspective of information seeking, attempts to facilitate the process and achieve high return on information in a ubiquitous environment. Beginning with a computational model to efficient management of user-generated contents, a progressive search algorithm is applied to obtain correlated information in accordance with user contexts. The above mentioned process is meanwhile normalized by a dynamic temporal-spatial model, which is primarily proposed to highlight the significance of timely information that matches the ultimate task of users at specific moment. Experiment results demonstrate the performance of proposed model and algorithm, and also the feasibility of this study.