Exploring the Algorithm for Diagnosing Burnout of Counselors in Higher Vocational Colleges in Guangdong Province Based on Emotion Recognition

Abstract

The main objective of the research is to deal with the drawbacks of conventional self-report tests by designing and validating a novel Emotion Recognition (ER) algorithm for assessing Counsellor Burnout (CB) in Higher Vocational Colleges (HVC) in Guangdong Province. Applying the Maslach Burnout Inventory (MBI) Survey Questionnaires, voice and facial expression files, and text analysis of counselling sessions, researchers collected an enormous dataset from 97 counsellors. The study employed a multimodal fusion layer to integrate multiple neural network architectures, namely RNNs for audio content, CNNs for video content, and a BERT+LSTM hybrid for textual analysis. This enabled a thorough examination of each ER. By dramatically exhibiting an accuracy of more than 92% for identifying the symptoms of burnout, this method is better than the conventional methods that have been used in previous years. The technique, which successfully recognizes complex, real-time emotions that are frequently ignored when individuals communicate emotions in person, can be used to obtain a more accurate and expert evaluation of burnout because of its capacity to recognize these reactions accurately. Different sources of data and contemporary Machine Learning (ML) methods have enhanced CB prediction and observation. This approach enhances counsellors' emotional stability and assists in developing pedagogical mental health interventions. Findings demonstrate that ML techniques may enhance mental assessment and counselling in demanding employment.