Location tracking technique for Regional ENF Classification Using ARIMA

Recently, a digital forensics technology has emerged that uses electrical network frequency (ENF) signals depending on the geographical environment for power supply. Through ENF data training, the technology finds the area where the signal occurred when any data was given. At this time, forecasts are usually made through interpolation based on trained data. In this paper, we proposed a location tracking method that does not require separate interpolation in the case of when trying to find the grid where the signal appears. The network frequency signal is collected from the streaming videos of the online multimedia services and the ENF signal is extracted using the secondary interpolation FFT (QIFFT) in the collected file. Subsequently, the extracted ENF signals are grouped into a constant size and trained through the Automatic Integrated Moving Average (ARIMA). Then, analyze the coefficients of regional ENF data to find meaningful sorting values in each region. This suggested how to track where online files are played and verified the accuracy of the predicted locations.