Data Mining Predictive Analysis For Perceived Stress Factors In Radiological Workers

ObjectiveExcessive perceived stress has a great influence on physical and mental health towards occupational groups in particular in radiological workers.The study of mental health is mainly about single element and there are defects in the conventional analytical methods.The study aims to show the interaction and the quantitative relationship towards psychological stress, life events in the crowd. It's laid the foundation for the improving physical as well as mental well-being, to provide new approach to support in mental health research by data mining.Subject and methodsThe data of this thesis was collected from all the medical radiological workers of the Institute of Occupational Disease Prevention of Henan province in April 2007. Apart from the general demographic surveys, the thesis has also collected psychology questionnaire of the Chinese version Scale Health Scale(CHQ-12),Eysenck Personality Questionnaire EPQ Chinese Version (EPQ-RSC),Chinese version of Perceived Stress Scale(CPSS),Maslach burnout inventory general version (MBI-GS),Life Event Scale (LES),Trait Coping Style Questionnaire (TCSQ),Social Support Rating Scale(SSS).To analyze the data by data mining, at first preprocessed the data to get the effective. The indicators compare with the total score about stress by correlation analysis and multiple regressions analysis. We selected the indicators which had a greater correlation with the stress, established the decision tree model to explore the communication between indicators and the stress.All data analysis realized in SPSS Clementine 11.1.Results1.According to analysis for correlation and multiple regressions, study shows that perceived stress effected jointly by emotional exhaustion, passive neglect, psychoticism, positive response and subjective support et al.We selected the indicators and established the decision tree model.2.We selected the arguments which has statistical significance for perceived stress as dependent variable by decision tree model,and we sorted grade arguments for eight variables, primary for emotional exhaustion, passive neglect, positive response, subjective support, psychoticism, positive events (good) to stimulate, negative response, the amount of professional efficacy.3.The rules for 11 can also analysed process and result clearly. It is not under perceived stress in (1)the scores of negative events (bad) to stimulate is less then 3, the amount of professional efficacy is less then 22, negative response is less then 25, positive events (good) to stimulate is less then 11,psychoticism is less then 3, subjective support is less then 28, positive coping is less then 34, passive neglect is less then 5,emotional exhaustion is less then 12; (2)the scores of subjective support is more then 28, positive coping is less then 34, passive neglect is less then 5,emotional exhaustion is less then 12; (3)the scores of positive coping is more then 34, passive neglect is less then 5,emotional exhaustion is less then 12.And it is under perceived stress in (1)the scores of negative response is more then 25, positive events (good) to stimulate is less then 11,psychoticism is less then 3, subjective support is less then 28, positive coping is less then 34, passive neglect is less then 5, emotional exhaustion is less then 12; (2)the scores of psychoticism is less then 3, positive events (good) to stimulate is more then 11,subjective support is less then 28, positive coping is less then 34, passive neglect is less then 5,emotional exhaustion is more then 12; (3)the scores of passive neglect is more then 5,emotional exhaustion is more then 12.4. The decision tree model predicts the outcome which is better by training and test sets, risk of statistics and consistency test.ConclusionsWe analysed the mental health in radiological workers and proposed the new idea to apply data mining technology to the mental health in radiological workers.It is quite adept at the psychological health of medical by data mining. We established the decision tree model by data mining and database. To analyse the stress and its relationship with their psychological health, social support,and to provide theory and scientific guidelines in radiological workers.