A study on the visual effect and user response of infomercials based on neural network analysis

This paper mainly takes the visual elements of infomercials as the perspective and the theory related to advertising effect as the basis, and uses the regression neural network model to study the influence of elements’ color, shape, brightness, etc. on the effect of infomercials and their functioning mechanism. A collaborative attention model combining the visual features of advertisement images and text features is constructed to improve the accuracy of users’ visual attention prediction. Element color, shape, brightness, etc., element position, size, number, etc., style selection and design all predicted social presence significantly (P=0.001), and the overall social facilitation effect and social presence predicted the advertising effect significantly, with the standardized coefficients of 0.617, 0.847, and 0.835, respectively, with a P- value equal to 0.001. Advertisement likability, advertisement aesthetics, and advertisement brand likability were negatively related to the average visual attention intensity. degree are negatively correlated with the average visual attention intensity, with correlation coefficients corresponding to -0.68, -0.86 and -0.84, respectively, which suggests that the more aesthetically pleasing the visual effect, the faster the user’s attention is perceived.