Integrating Social Media Insights for Innovation Performance Enhancement: A Transformer-Based Analysis

With the development of the times, social media in daily work is increasingly popular, which has a potential impact on employee innovation performance. However, current research only focuses on single factor analysis. In order to enhance this research, this paper studies transformer model, attention mechanism, and other advanced data analysis methods to investigate the subtle relationship between employees’ use of social media and its impact on a company’s innovation performance. Our research introduces a multidimensional approach that encompasses both organizational and individual perspectives, offering a comprehensive understanding of social media’s dualistic nature in fostering innovation. Employing an innovative methodological framework, we integrate factor analysis with a transformer-based feature fusion module, effectively capturing and analyzing the rich semantic nuances embedded in employee-generated social media content. This approach allows for the extraction and synthesis of key sentiment indicators, facilitating a more granular analysis of how social media engagement influences innovation-related behaviors and outcomes. Findings reveal that the strategic use of social media within corporate environments can significantly enhance innovation performance by providing a fertile ground for knowledge sharing, collaborative engagement, and the nurturing of a culture conducive to innovation. The proposed multimodal data fusion technique demonstrates superior accuracy in sentiment analysis, surpassing traditional unimodal approaches by significant margins. These insights contribute to the academic discourse on technology management and knowledge economy and offer practical implications for organizations aiming to harness social media’s potential in augmenting their innovation ecosystem.