Automation of Stress Recognition Using Subjective or Objective Measures

Automation of stress recognition involves collecting stress-reflecting data and then using a computational technique to estimate the stress level corresponding to that data. With the advent of objective sensors or mobile context measurement, automated stress recognition revolves around and is restricted to only these measures. Traditional subjective questionnaires and self-reports, though highly reliable, have been bounded to only clinical settings, without extending their scope to general population and natural environments. This review discusses some widely used questionnaires for measuring stress which can prove to be equally efficient to their conventional paper-based application when computerized. A summary on highly accepted objective measures of stress, including physiological, behavioural and contextual data, is provided next. Past research work using these subjective and objective measures for automatically recognizing stress is also reviewed. A comparative analysis of questionnaires and objective measures is conducted, which helped in highlighting significant advantages and disadvantages of the measures. Use of reports or questionnaires for automating stress recognition is suggested, owing to improved costs, high reliability and user trust.