Abstract—This paper presents anapproach of hybridizing two or more artificial intelligence (AI) techniques which arebeing used to fuzzify the workstress level ranking and categorize the rating accordingly. The use of two or more techniques (hybrid approach) has been considered in this case, as combining different techniques may lead to neutralizing each other’s weaknesses generating a superior hybrid solution. Recent researches have shown that there is a need for a more valid and reliable tools, for assessing work stress. Thus artificial intelligence techniques have been applied in this instance to provide a solution to a psychological application. An overview about the novel and autonomous interactive model for analysing work-stress that has been developedusing multi-agent systems is also presented in this paper. The establishment of the intelligent multi-agent decision analyser (IMADA) using hybridized technique of neural networks and fuzzy logic within the multi-agent based framework is also described. Keywords—Fuzzy logic, intelligent agent, multi-agent systems, neural network, workplace stress. I. INTRODUCTION ORK stress has become a widespread concern in Australiaand other countries. Work related stress affects people from all professions and is a growing concern in Australia and overseas, as it is reported as a common cause of occupational illness. Workplace stress has been defined as a pattern of emotional, cognitive, behavioral and psychological reactions to adverse and noxious aspects of the work content, the organisation and the work environment. Stress is also defined as an adverse reaction people experience as a result of pressure at work from job demand, harassment and injustice in the work environment [1]. Thus there is a need for a psychological risk assessment system that can be used by individual user or users from organizations. In order to assess such psychological risk arising with the workplace, an autonomous intelligent agent based system is currently being developed [2]. In the first stage of the system development, the system allows user to take online work stress related survey based on a particular industry and generate a feedback in real time comparing and benchmarking user result against national benchmark data [2], [3]. However it has been found that as more and more users complete the survey, the data for benchmarking needs A. Ghosh is a PhD student in the School of Electrical and Information Engineering, University of South Australia, Mawson Lakes, Adelaide South Australia (phone: 0618-8302 3649; e-mail: Anusua. Ghosh@ unisa.edu.au). A. Nafalski is a professor of electrical engineering in the School of Electrical and Information Engineering, University of South Australia (phone: 0618-8302 3932;-e-mail: Andrew.Nafalski@unisa.edu.au). J. Tweedale, PhD., is with the Air Operation Division, Defence Science and Technology Organisation, Edinburgh, Adelaide, South Australia (e-mail: Jeffery.Tweedale@dsto.defence.gov.au). M. Dollard is a professor of psychology and Director for the Centre for Applied Psychological Research, University of South Australia, Adelaide (e- mail: Maureen.Dollard@unisa.edu.au). updating for accurate comparisons. This has been achieved in the second stage by using multi-agents technology, for effectively updating and maintaining both the user and the benchmarking databases based on a threshold value [3]. Analyzing work stress related data collected using the survey tool forms the important part of the system, as the data need to be preprocessed, analysed, benchmarked and generate feedback to the user in real time. Thus the decision process has been transformed to include multi-agent and a new decision analyzer. The next phase of the development of the system is based on research that integrating two or more techniques to solve a complex problem usually optimizes the performance of the system. This paper presents an insight into the intelligent multi-agent decision analyzer (IMADA) which is developed by integrating neural network with fuzzy logic. In section II an autonomous intelligent interactive system is discussed, then artificial neural network in section III. Multi- agent system in section IV is followed by the design and methodology of IMADA in section V and finally conclusion and future works are presented in section VI. II. AUTONOMOUS INTELLIGENT INTERACTIVE SYSTEM The autonomous interactive stress assessment system is developed using AI techniques applying Australian workplace barometer (AWB) tool and the StressCafé e-portal, which are explained below: A. Australian Workplace Barometer The Australian workplace barometer (AWB) is a tool designed to measure workplace psychological risk in relation to health and work outcome. It is a world class national survey aimed to assess work-related stress that give rise to psychological factors that generally affect people's well-being and their work [4]. B. StressCafe - An Interactive Website The StressCafé [5] is an interactive website that is the single point of contact for measuring work stress, generating feedback, sharing information, and bench-marking psychosocial hazards in the Australian workplace [5]. Within the StressCafé using the AWB online an interactive, autonomous and intelligent system is implemented. The autonomous interactive system within StressCafé provides feedback to participants who complete a work-based psychosocial risk assessment survey by comparing individual results to AWB benchmark scores. As such, the StressCafé enables workers, employers, researchers and academics to access AWB online evidence-based psychosocial risk assessment tools which can be used to extrapolate individual/organizational level data and then compare, and benchmark nationally to provide an immediate feedback. Anusua Ghosh, Andrew Nafalski, Jeffery Tweedale, and Maureen Dollard Hybridized Technique to Analyze Workstress Related Data via the StressCafé W World Academy of Science, Engineering and Technology International Journal of Computer and Information Engineering Vol:6, No:10, 2012 1266 International Scholarly and Scientific Research & Innovation 6(10) 2012 scholar.waset.org/1307-6892/8998 International Science Index, Computer and Information Engineering Vol:6, No:10, 2012 waset.org/Publication/8998