Accurate sales forecasting of fresh food is imperative for retailers. It facilitates maintaining optimal inventory levels, thereby enhancing customer satisfaction, boosting revenue, and minimizing waste. However, sales sequences of fresh food are subject to multiple compounded factors, exhibiting nonlinearity and non-stationarity, posing challenges for prediction. This paper proposes a novel multi-variable hybrid model, VMD-LSTM-BMA, based on variational mode decomposition (VMD), long short-term memory (LSTM) neural networks, and Bayesian model averaging (BMA), for daily fresh food sales forecasting. Utilizing the posterior distribution generated by BMA, we calculate prediction intervals at various confidence levels to quantify the uncertainty of the forecasting outcomes. Employing a daily banana sales dataset from a retail chain supermarket, we validate the predictive performance of the proposed hybrid model at different aggregation levels. The results demonstrate that our VMD-LSTM-BMA framework achieves superior point forecasting accuracy compared to other models. In most instances, the prediction intervals provided by VMD-LSTM-BMA exhibit a higher prediction interval coverage probability (PICP) and a narrower interval width. Our proposed hybrid model operates robustly and efficiently, capable of providing reliable guidance for retailers’ replenishment and ordering processes, thereby mitigating the risks of out-of-stock and excess inventory.

Goal contents pursuit serves as an effective indicator to predict individual behaviors and life satisfaction. In current studies, the measures of self-reports are limited by the expenses and subjectivity, Despite the fact that machine learning and deep learning models can effectively utilize user data from social media to objectively classify goal contents pursuit, the workload associated with supervised learning is substantial. By using appropriate prompt patterns, Large Language Models (LLM) can execute diverse tasks in a prompt manner. Thus, we develop six prompt patterns for classifying goal content pursuits of users on social media. The results indicate that some of these prompt patterns successfully classify intrinsic and extrinsic goals. Overall, this research presents an usable prompting paradigm and approach, PsyPrompt, which enhances the methodology of objectively classifying individual goal content pursuits using LLMs.

Public opinion plays a pivotal role in politics, influencing political leaders' decisions, shaping election outcomes, and impacting policy-making processes. In today's digital age, the abundance of political discourse available on social media platforms has become an invaluable resource for analyzing public opinion. This paper focuses on the task of detecting political stances in the context of the 2020 US presidential election. To facilitate this research, we curate a substantial dataset sourced from Twitter, annotated using hashtags as indicators of political polarity. In our approach, we construct a bipartite graph that explicitly models user-tweet interactions, which provides a comprehensive contextual understanding of the election. To effectively leverage the wealth of user behavioral information encoded in this graph, we adopt graph convolution and introduce a novel skip aggregation mechanism. This mechanism enables tweet nodes to aggregate information from their second-order neighbors, which are also tweet nodes due to the graph's bipartite nature. Our experimental results demonstrate that our proposed model outperforms a range of competitive baseline models. Furthermore, our in-depth analyses highlight the importance of user behavioral information and the effectiveness of skip aggregation.

Given the development and abundance of social media, studying the stance of social media users is a challenging and pressing issue. Social media users express their stance by posting tweets and retweeting. Therefore, the homogeneous relationship between users and the heterogeneous relationship between users and tweets are relevant for the stance detection task. Recently, graph neural networks (GNNs) have developed rapidly and have been applied to social media research. In this paper, we crawl a large-scale dataset of the 2020 US presidential election and automatically label all users by manually tagged hashtags. Subsequently, we propose a bipartite graph neural network model, DoubleH, which aims to better utilize homogeneous and heterogeneous information in user stance detection tasks. Specifically, we first construct a bipartite graph based on posting and retweeting relations for two kinds of nodes, including users and tweets. We then iteratively update the node's representation by extracting and separately processing heterogeneous and homogeneous information in the node's neighbors. Finally, the representations of user nodes are used for user stance classification. Experimental results show that DoubleH outperforms the state-of-the-art methods on popular benchmarks. Further analysis illustrates the model's utilization of information and demonstrates stability and efficiency at different numbers of layers.

It is convenient for people to take notes directly on book pages while reading. But for public books, such as those borrowed from library, direct marking is inappropriate and forbidden. To solve this problem, we propose D-Note, a new computer-aided digital note taking system. D-note registers and distinguishes book page according to its visual features. It recognizes hand interactions on page with a depth image from Kinect. Meanwhile it gets user operation intentions by speech recognition and dynamically creates a page-related digital note. The experiments and user study show that D-note is a worthwhile way to protect books and a good reading assist.

We describe a user interface and interaction technique, named ‘Javelin’, designed for large display environments. It provides quick access to random screen regions and manipulation methods for screen widgets which are difficult or impossible to reach. It consists of a dynamic global thumbnail, a touchpad widget that drives the screen cursor, and a teleport widget in which interactions are transferred to its target screen region. Javelin can be easily integrated into many programs to optimize their interaction performance in large screens. The experiment and user study show that Javelin can extend user access field and enhance widget manipulation in large displays.