The following is an example of the implementation of keystroke dynamics based on the dataset described in the study Comparing anomaly-detection for keystroke dynamics mentioned in the previous section; the dataset is also available for download in .csv format at https://www.cs.cmu.edu/~keystroke/DSL-StrongPasswordData.csv.
As anticipated, the dataset consists of 51 subjects, each typing 400 passwords; also among the measures collected are these hold times (represented in the dataset with the label H):
- Keydown-keydown time (labeled as DD)
- Keyup-keydown time (labeled as UD)
The code for the keystrokes detection script is as follows:
import numpy as npimport pandas as pdfrom matplotlib import pyplot as plt ...
