Prediction of purchase decision-making on digital live streaming Interface on MT: A comparative research of multi-modal Human-Machine Interaction

Abstract

In the emerging live-stream shopping, evaluating and establishing the best interface design for live streaming platform to improve user experience and effect transactions is essential. Nevertheless, traditional methods such as subjective interviews and task-observation have limitations to assess the neural mechanisms involved in this process. This study sought to apply neuroscience methods in analyzing and evaluate 3 different live streaming APP’s Human-machine Interaction with users and further use Support vector machine (SVM) classifiers based on different combination of users’ electrodermal activity, electroencephalogram and eye movement to predict users’ purchase decision. To this end, 35 participants were required to complete 4 shopping tasks on 3 live streaming APP respectably, and their multi-dimensional subjective and objective data were collected and analyzed. The results showed that the tonic SC was significantly smallest in the A APP than other two APP, and the B APP had smallest fixation frequency, implying that B App might need less effort in interaction with users and obtained more positive user experience. Moreover,the fusion of eye movements, EDA and EEG data could fairly improve the performance of design decision making classification. Then, it is possible that applying multi-modal physiological data synchronization would be an effective approach to assess live streaming platform and further improve the integration quality, or develop advanced tailor-made product recommendation system.

Keywords: Live Streaming APP, Electrodermal Activity, Electroencephalogram, Eye Movement, Support Vector Machine, Purchase Decision, making

DOI: 10.54941/ahfe1003439