Sarcasm detection is a critical aspect of sentiment analysis on social media platforms, where expressions often convey meanings that deviate from their literal interpretations. Although Chinese is considered a high-resource language, detecting sarcasm remains challenging due to the absence of large-scale, annotated datasets and sarcasm’s complex, context-dependent, and culturally nuanced nature. This paper presents a multi-dimensional approach to sarcasm detection in Chinese, effectively addressing these challenges. We introduce Chinese TikTok Sarcasm Dataset (CTSD), a relatively large-scale, manually annotated collection of sarcastic comments sourced from TikTok, to address data scarcity. We demonstrate the effectiveness of the new dataset through extensive experiments using state-of-the-art models and introduce a hybrid deep learning model, BERTWWM-CNN-BiLSTM, which combines the power of Bidirectional Encoder Representations from Transformers with whole word masking (BERT-wwm) for rich contextual embeddings, Convolutional Neural Network (CNN) to capture local n-gram features, and Bidirectional Long Short-Term Memory network (BiLSTM) for learning long-range dependencies. Experimental results show that BERTWWM-CNN-BiLSTM outperforms baseline models, achieving an F1-score of 79.07% and an Area Under the ROC Curve (AUC) of 85.51%, reflecting improvements of 3.58% and 10.63%, respectively. These findings highlight the potential of our approach in enhancing sarcasm detection in social media text.

Traditional recommender systems are dedicated to recommending items of interest to consumers, but consumers are more concerned about how to purchase optimal bundles (combinations of items) under the constraints of their limited budgets. In this paper, an approach for recommending bundles according to users’ budgets is proposed based on Graph Convolutional Networks (GCN). The GCN propagates over a heterogeneous graph consisting of the user-bundle interactions, user-item interactions, and bundle-item affiliations. The price attributes of bundles and items are also considered to capture users’ preferences at two levels: the price of a bundle as a whole and the price of individual items contained in the bundle. Besides Hit Rate (HR) and Normalized Discounted Cumulative Gain (NDCG), a metric that takes into account the similarity between a recommended bundle and the user’s target bundle is designed to evaluate the performance of our model. Experimental results show that the model achieves state-of-the-art performance on three datasets. The approach of recommending bundles according to users’ budgets is more in line with the theory of consumer choice, which provides a fresh perspective for the research on recommender systems.