Target Audience Identification for Products in E-commerce Live Streaming

Over the last several years, live streaming on e-commerce platforms has emerged as an important tool for companies and online personalities to connect with a diverse audience. However, identifying the optimal audience for a certain product in real time remains a difficult issue. It is critical for tailored marketing and efficient content delivery to categorize viewers into distinct groups (for example, youth, middle-aged and old, men, women, special groups, and the general audience). The purpose of this work is to introduce a new technique for identifying target audiences in live e-commerce data using cue-based learning. Experimental results reveal that the proposed cue-based strategy outperforms standard classification models and pre-trained models (for example, BERT), with considerable increases in accuracy, precision, recall, and F1 score.