Thesis:

Comparison of Multiple Regression and Artificial Neural Network in Predicting Mental Burnout Based on Personality Traits and Coping Styles

Mental exhaustion is a major psychological health challenge in women, the accurate modeling of which requires efficient linear and nonlinear analysis tools. The present study aimed to compare the power and accuracy of multiple linear regression with the Multi-Layer Perceptron Artificial Neural Network (ANN-MLP) in predicting mental exhaustion based on the Big Five personality traits and coping styles in women. The research design was descriptive-correlational and predictive-modeling. The statistical population consisted of women, from whom 384 individuals were selected via available sampling and completed the Maslach Burnout Inventory (MBI), the NEO Five-Factor Inventory (NEO-FFI), and the Lazarus and Folkman Ways of Coping Questionnaire (WOCQ). After preprocessing, the data were divided into training (80%) and testing (20%) sets. Modeling was implemented using multiple regression and an ANN-MLP with an 8-10-1 structure, and their performance was compared based on R^2, RMSE, and MAE indices. The regression analysis results showed that neuroticism and avoidant coping style directly and significantly, and conscientiousness, extraversion, agreeableness, and problem-focused coping style inversely and significantly predict mental exhaustion (p<0.01). Furthermore, on the test data, the artificial neural network model demonstrated significantly higher accuracy and generalizability with a higher coefficient of determination (R^2=0.81) and lower errors ("RMSE"=3.82 and "MAE"=2.94) compared to linear regression (R^2=0.69, "RMSE"=4.95, and "MAE"=3.88). Feature importance analysis also identified neuroticism, avoidant coping, and problem-focused coping as the most important variables. Accordingly, the relationships of personality and coping variables with mental exhaustion also follow complex nonlinear patterns, and the application of artificial intelligence computational models can serve as a precise tool in screening and designing psychological interventions for women.

Keywords: Mental exhaustion, Big Five personality traits, Coping styles, Linear regression, Artificial neural