Abstract: Digital human live technology can inject new vitality into the live industry and promote the development of digital human live streaming. However, affected by multiple factors, digital human live streaming in the era of weak artificial intelligence has many problems such as poor flexibility and insufficient emotional cognition, resulting in a serious lack of user participation. How to improve the quality of digital human live streaming and enhance user participation is an urgent management problem to be solved. However, the existing research on the mechanism of the influence of digital human live streaming user participation behavior is insufficient, especially the lack of in-depth research on the influence of different factor groupings on user participation behavior.
This study constructs a model of the influence mechanism of digital human live streaming user participation behavior based on the S-O-R theory. In terms of external stimuli, it introduces impression cues, interactive cues and emphatic cues from the anthropomorphic cues perspective, and entertainment atmosphere and interpretation atmosphere from the live streaming atmosphere perspective. In terms of intrinsic organism, it introduces trust and social presence based on the social exchange theory and the social presence theory. In terms of individual behavioral response, it introduces onlooker participation behavior, expressive participation behavior, and purchasing participation behavior according to the degree of user involvement. Then the study collects 329 valid data related to the user participation behavior of digital human live streaming through questionnaire research, and identifies the key influencing factors and grouping patterns of the user participation behavior path at the single and holistic levels by applying structural equation modeling (SEM) and fuzzy set qualitative comparative analysis (fsQCA), respectively.
The results of the study are as follows: (1) Interactive cues and interpretation atmosphere have a significant positive effect on perceived trust, with interpretation atmosphere exerting a stronger effect; Impression cues, emphatic cues, entertainment atmosphere and interpretation atmosphere have a significant positive effect on social presence, with entertainment atmosphere exerting a stronger effect; social presence has a significant positive effect on perceived trust. Both perceived trust and social presence have a significant positive effect on users' engagement behaviors. However, none of these influencing factors can constitute a necessary condition for user participation behavior, and a variety of factors are needed to play a synergistic role. (2) High entertainment atmosphere, high perceived trust and high social presence are present in all configurations of onlooker participation behavior, expressive participation behavior, and purchasing participation behavior; Impression cues, empathic cues, and interpretation atmospheres cross over as core conditions in most groupings. Based on the cross-occurrence situation, the onlooker participation behavior is emotional content-driven and image content-driven, the expression participation behavior is emotion-driven, and the purchasing engagement behavior is emotional image-driven and emotional social-driven.
Compared with existing research, this study makes important breakthrough in the following two aspects: For one thing, the virtual character identity of digital human live streaming leads it to be different from traditional live streaming, and the mechanism of user participation behavior is also different. The variables introduced in this study from the perspective of anthropomorphic cues can effectively reflect the differences between traditional live streaming and digital human live streaming, and the research on the influencing mechanism of the behavior of digital human live streaming users' participation has a certain degree of foresight. For another, the generation of user participation behavior of digital human live streaming is the result of the combined effect of many factors. Complementary and combinational relationships between antecedent variables can be recognized from a group perspective, and excavate the internal mechanism of user participation behavior at a deeper level, so that the research is more comprehensive.
To a certain extent, this study reveals the intrinsic influence mechanism generated by the user participation behavior of digital human live streaming, which can provide guidance for relevant practitioners. It can not only provide operation strategies and optimization suggestions for digital human live streaming technology developers and live streaming platforms, but also promote healthy and sustainable development of the digital human live streaming industry, and accelerate the release of the new kinetic energy of the digital economy.

Key words: Digital Human Live Streaming, S-O-R Theory, User Engagement Behavior, Anthropomorphic Cues, Live Streaming Atmosphere

摘要： 数字人技术可为直播行业注入新活力,如何提升数字人直播质量,增强用户参与度是亟待解决的管理问题。本研究基于S-O-R理论构建数字人直播用户参与行为影响机制模型,通过结构方程模型(SEM)和模糊集定性比较分析(fsQCA)识别用户参与行为关键影响因素及组态路径。研究发现:(1)交互型线索和解说氛围正向影响信任;印象型线索、移情性线索、娱乐氛围和解说氛围正向影响社会临场感;社会临场感正向影响信任,信任和社会临场感正向影响用户参与行为;所有影响因素均不能构成用户参与行为的必要条件,需多种因素协同发挥作用。(2)高娱乐氛围、高信任和高社会临场感均存在于围观式、表达式及购买式参与行为的所有组态中,对用户参与行为具有促进作用。在此基础上,围观式参与行为产生路径为情感内容驱动路径及形象内容驱动路径;表达式参与行为产生路径为情感驱动路径;购买式参与行为产生路径为情感形象驱动路径和情感社交驱动路径。

关键词: 数字人直播, S-O-R理论, 用户参与行为, 拟人化线索, 直播氛围