A better positioning with BLE tag by RSSI compensation through crowd density estimation
The paper aims to interpret visitor behaviors by analyzing a large scale Wi-Fi location data obtained at Interop Tokyo 2014. We first detected a situation of a visitor's stay at a booth, calculated the sojourn time of the visitor and represented each visitor as a booth vector whose element value is the total sojourn time of the visitor at a booth. Then, we classified the visitors by k-Means. By using category labels of each booth, we analyzed characteristics of clusters. The results illustrate that some categories appear at the top rank of most clusters, and the area of the booths included in each cluster is mostly some specific small one, not the entire one.