Overview This dataset contains the power consumption and ground truth for 40 runs of 4 activity hours on an intel NUC micro-PC. The power consumption is pre-processed from 10 KSPS to 20 SPS with a median filter. The ground truth contains the label for each sample with: -1=UNKNOWN, 0=OFF, 1=IDLE, 2=HIGH LOAD and 3=REBOOT. Usage To use the npy format, load the python package numpy and load the data with the `load` function: ```import numpy as np trace = np.load("trace_4.npy")gt = np.load("gt_12.npy")``` The CSV format can be loaded in a variety of software. Changelog An index issue was present in the first version. The ground truth and trace were offset by 1 (trace n was associated with ground truth n-1). Moreover, trace 2 did not have a corresponding ground truth. This new version fixes this issue and removes the trace without ground truth. Naming convention was also changed to include leading 0 and make listing easier in file explorers.