Accelerometer-Based Evaluation of the Human Psychological State While Viewing Content on Smartphones

Abstract

In recent years, with the abundant availability of digital content, smartphone users have often faced difficulty finding what they prefer. Several recommendation systems have been developed, but conventional ones do not reflect user evaluation. This study focuses on invisible and feeble physiological tremors to reflect user evaluation. The acceleration data of three subjects while reading comic books on a smartphone were collected by using the 3D accelerometer of a smartphone and a smartwatch. Then, the acceleration data in front of the wrist were obtained from the difference values of the two devices. In this method, the measurement was not affected by posture. As a result of performing fast Fourier transforms, the amplitude peak was observed between 2–4 Hz in each data item. In general, the main frequency of a physiological hand tremor is 6 Hz and it is said to decrease as the mass increases. For this reason, it was suggested that the acquisition data included physiological tremors in front of the wrist. Moreover, there was a negative correlation between the user interest in each story and the integral value of amplitude. In particular, the difference value shows a strong correlation. Accordingly, using the difference values of the two devices enables a more accurate evaluation of the psychological state.