Abstract
In this report, the authors propose a neural network model to estimate anaerobic threshold for Emotional Fitness, exercising for mental health care and improvement. First, subjects run for ten minutes, increasing pace from 4[km/h] to 8[km/h] in every 30[sec] in increments of 0.1[km/h] on a treadmill. The heart rate and the two dimension mood scale (TDMS) are measured before and after exercise to record level of comfort. We approximate relation between running pace and the heart rate squared by a sigmoid function. Next, we construct a neural network model with inputs of selected heart rate and subject's exercise history and output of AT. After neural network training, the proposed model is validated using unknown data.
