MUFG Bank (United Kingdom)

By analysing students' behavioural states, one can evaluate their level of participation, engagement and focus in learning activities. Therefore, this study designs a method for identifying student behaviour states in business English listening classes based on the SSD algorithm. The behaviour state features are input into the improved SSD algorithm model. Within the model, behavioural state features are gradually extracted through operations such as convolution and pooling, and the classification and recognition results of student behavioural states are output. In the experimental results, the Cohen's Kappa coefficient of the identification results obtained by this method can reach 0.974, and the global minimum and maximum values of the Matthews correlation coefficient are 0.80 and 0.95, indicating that the identification results of this method are effective.