KeyGAN: Synthetic keystroke data generation in the context of digital phenotyping

Alejandro Acien, Aythami Morales, Luca Giancardo, Rubén Vera-Rodríguez, Ashley A Holmes, Julian Fiérrez, Teresa Arroyo-Gallego

Abstract: Highlights•A benchmark method for real-world keystroke data generation: KeyGAN.•This method generates not only realistic but also diverse synthetic keystroke patterns.•Deep experimentation to test KeyGAN synthesizers based on referee evaluation methods.•The synthetic keystroke data also simulates typing patterns of Parkinson’s disease patients.•The synthetic data can decrease the required number of real-world data, without compromising the model’s performance.