Search results

Short-form video advertisements have recently gained popularity and are widely used. However, creating attractive short video advertisements remains a challenge for sellers. Based on the visual-audio perspective and signaling theory, this study investigated the impacts of three visual features (number of shots, pixel-level image complexity and vertical versus horizontal formats) and two audio features (speech rate and average spectral centroid) on user engagement behavior.

We conducted a field study on TikTok. To test our various hypotheses, we used regression analysis on 2,511 videos containing product promotion information posted by 60 sellers between January 1, 2020 and November 20, 2021.

For visual variables, the number of shots and pixel-level image complexity were found to have nonlinear (inverted U-shaped) relationships with user engagement behavior. The vertical video form was found to have a positive effect on comments and shares. In the case of audio variables, speech rate was found to have a significant positive effect on shares but not on likes and comments. The average spectral centroid was found to have significant negative influences on likes and comments.

This study provides specific suggestions for sellers who create short-form videos to improve user engagement behavior.

This study contributes to the literature on short-form video advertising by extending the potential drivers of user engagement behavior. Additionally, from a methodological perspective, it contributes to the literature by using computer vision and speech-processing techniques to analyze user behavior in a video-related context, effectively overcoming the limitations of the widely adopted survey method.

This paper addresses the limitations of current graph neural network-based recommendation systems, which often neglect the integration of side information and the modeling of complex high-order interactions among nodes. The research motivation stems from the need to enhance recommendation performance by effectively utilizing all available data. We propose a novel method called MSHCN, which leverages hypergraph neural networks to integrate side information and model complex interactions, thereby improving user and item representations.

The MSHCN method employs a hypergraph structure to incorporate various types of side information, including social relationships among users and item attributes, which are essential for enriching user and item representations. The k-means clustering algorithm is utilized to create item-associated hypergraphs, while sentiment analysis on user reviews refines the modeling of user interests. Additionally, hypergraphs are constructed for user-user and item-item interactions based on interaction similarity. MSHCN also incorporates contrastive learning as an auxiliary task to enhance the representation learning process.

Extensive experiments demonstrate that MSHCN significantly outperforms existing recommendation models, particularly in its ability to capture and utilize side information and high-order interactions. This results in superior user and item representations and improved recommendation performance.

The novelty of MSHCN lies in its use of a hypergraph structure to integrate diverse side information and model intricate high-order interactions. The incorporation of contrastive learning as an auxiliary task sets it apart from other hypergraph-based models, providing a significant enhancement in recommendation accuracy.

To reduce the difficulty of the sewing process and promote the automation process of fabric sewing, a soft finger-assisted feeding method is proposed to investigate the effect of sewing process parameters on the quality of automatic sewing.

Taking cotton woven fabrics as an example, the causes of sewing deviation are firstly investigated from three aspects: fabric properties, sewing speed and sewing edge position. By simulating the sewing action of human hands, the method of reducing sewing deviation by using soft fingers to press and feed the fabric is proposed. Then, four sewing process factors, namely, robot arm end pressure, sewing machine speed, sewing needle gauge and stitch density, were selected, and three levels were set for each factor to design orthogonal sewing experiments. The sewing deviation of 1# sample under different sewing processes was measured, and the optimal parameter matching for automatic sewing of this specimen was derived.

The findings demonstrate that, while sewing cloth automatically, the sewing deviation is significantly influenced by the robotic arm's end pressure, sewing speed, and stitch density, whereas the sewing deviation is not significantly impacted by the needle number.

The findings offer fundamental information for the development of an automated sewing procedure using soft fingers, which has theoretical and real-world application value to speed up the intelligent modernization and transformation of the apparel industry.