The rapid growth of electronic waste around the world has led to increased recycling problems. With the development of information technology, e-commerce has become a new trend in electronic waste recycling. This research designs two experiments to study the effect mechanism of green information on participation intention (PI) for online recycling websites. We found that providing green information can increase the PI of consumers, including environmental knowledge of product recycling (EKPR) and environmental corporate social responsibility (ECSR). Green perceived value (GPV) and green trust (GT) play mediating roles on the impact of green information on PI. In addition, the recycling platform reputation (RPR) plays a moderating role in the effect of ECSR on GT. Theoretical and managerial implications, along with avenues for future research, are discussed.

E-waste online recycling is a new recycling model that has received strong support from governments worldwide. It integrates multiple processes such as online inquiry and offline delivery. The complex online–offline interaction makes residents’ willingness to participate susceptible to external factors, hindering their behavior. To better transform residents’ recycling intentions into behavior, this paper primarily uses a questionnaire survey to explore the intention–behavior gap in e-waste online recycling participating. Our findings show that residents’ e-waste online recycling intention directly predicts their behavior, but the implementation plan plays a partial mediating role between them. Additionally, perceived value and platform interactivity moderate the impact of e-waste online recycling intention towards participation behavior. Subsequent in-depth interview analysis further also confirmed these findings. The study contributes to understanding residents’ online recycling behaviors and fostering the sustainable development of this model.

With the development of cognitive computing and intelligent media communication, recognizing the user’s emotion using electroencephalography (EEG) has garnered increasing attention. However, building a cross-subject emotion recognition model with good generalization performance is really difficult. To achieve remarkable performance, a label-refined domain adversarial neural network is designed and EEG samples are subdivided into 6 groups according to the differences in users' brain activities under different movie stimuli. Based on the experiment of SEED and SEED-IV datasets, the classification accuracy of our method reaches 93.35±4.14% and 84.08±6.01%, achieving the state-of-the-art performance. Experiments show that using different movie stimuli to refine labels can achieve efficient cross-subject emotion recognition.