Abstract: [Purpose/Significance] In the context of the digital age, knowledge collaboration platforms such as online Q&A communities, academic forums, and various professional networking platforms have become important venues for knowledge sharing and collective wisdom. These platforms bring together users from different fields, with diverse professional backgrounds and levels of expertise. They actively engage in problem solving, exchange views, and form complex and dynamic social networks. Online knowledge collaboration platforms not only enhance the accessibility of knowledge but also serve as incubators for interdisciplinary communication, problem solving, and innovative thinking by harnessing the collective wisdom and expertise of individuals. This article explores how to optimize the network structure of online knowledge collaboration platforms and balance the internal knowledge and expertise within teams. The goal is to promote cross-domain information flow, prevent the formation of information silos, and promote the creation, dissemination, and application of knowledge through collective knowledge collaboration. [Methods/Process] Due to the diversity of participants' backgrounds, experiences, and viewpoints, effectively managing and coordinating this heterogeneity becomes a critical issue. Additionally, the quality and efficiency of knowledge collaboration is also influenced by the characteristics of the network structure, such as the flow of information paths, the role of key nodes, and the interaction patterns of small groups. This study is based on actual data from Stack Overflow, the world's largest programming Q&A website. It focuses specifically on the following aspects of influence: clustering coefficient, node centrality, edge span, user knowledge heterogeneity, and user experience heterogeneity. By constructing a negative binomial regression model, the study investigates how network structure characteristics and team user heterogeneity affect the quality and efficiency of knowledge collaboration. [Results/Conclusions] The results show that, with respect to network structural characteristics, node centrality significantly improves the quality and efficiency of collaboration, and higher aggregation coefficients and larger span of connecting edges restrict information flow and are detrimental to the efficiency of knowledge collaboration. In terms of user heterogeneity, high heterogeneity in knowledge background and registration duration usually hinders collaboration, heterogeneity in experience heterogeneity in registration duration negatively affects collaboration effectiveness in both cases, heterogeneity in response acceptance rate only negatively affects collaboration quality, while heterogeneity in activity intensity positively affects it. In addition, this study still has shortcomings that deserve further exploration. First, future research could consider expanding the sample to include more questions on different topics and domains to increase the reliability and generalizability of the findings. Second, future research could focus on the dynamic changes of network structure and heterogeneity in order to better understand the impact of network structure on knowledge collaboration and to improve the prediction ability of collaboration effects; it could explore more deeply how different types of heterogeneity affect collaboration dynamics over time.

Key words: knowledge collaboration outcomes, social network structure, user heterogeneity, information behavior