Implementation of context prediction system based on event recurrence time

This study proposes a context prediction system considering event recurrence time for the accuracy improvement of prediction results. The system identifies current user’s movement through Beacon’s RSSI signal intensity existing in the user’s movement path. Subsequently, the event recurrence time required to move from the current position to the next position is calculated on the basis of cumulative data on temporal and spatial movement paths collected from a smartphone. The system utilizes the data as input data for the prediction of the next position. This study built a test bed to assess the proposed system’s prediction, collected user’s real time movement path data for 60 days, and compared the context prediction results with those of the two types of control groups. As a result of the experiment, the prediction accuracy of the control group A that used only user’s past position information was 86.97 %, and that of the control group B, which simultaneously used past position and time information, was 82.17 %. The prediction accuracy of the system proposed in this study that considered user’s past position information and event recurrence time was 87.21 %. Such a result was evaluated as improving prediction result by 8–16 %, compared with the control systems.