Clustering Mobile Traffic Data with Autoencoder Using Time-Series Encoded as Images.

With each new generation of mobile communication, the mobile traffic data are increasing exponentially. However, there are limitations to the capacity of base stations. As a result, it become necessary to manage mobile traffic efficiently. It will result in the understanding of the characteristics of mobile traffic, yet it is challenging to find research based on base station traffic volume. In this paper, we encode mobile traffic data of base station from multiple locations to image, and cluster the encoded images using autoencoder. Based on the labeled locations of several base stations we directly investigated, we confirm whether the unlabeled base stations are well grouped into shopping malls, offices, residential areas, and outside areas.