An increase in the carbon dioxide (CO₂) concentration within a vehicle can lead to a decrease in air quality, resulting in numerous adverse effects on the human body. Therefore, it is very important to know the in-vehicle CO₂ concentration level and to accurately predict a concentration change. The purpose of this research is to investigate in-vehicle concentration levels of CO₂, comparing the accuracy of an autoregressive integrated moving average (ARIMA) model and a long short-term memory (LSTM) model in predicting the change in CO₂ concentration. We conducted a field test to obtain in-vehicle original concentration data of CO₂ while driving, establishing a prediction model of CO₂ concentration with ARIMA and LSTM. We selected mean absolute percentage error (MAPE) and root mean squared error (RMSE) as the evaluation indicators. The findings indicate the following: (1) With the vehicle windows closed and recirculation ventilation mode activated, in-vehicle CO₂ concentration increases rapidly. During testing, CO₂ accumulation rates were measured at 1.43 ppm/s for one occupant and 3.52 ppm/s for three occupants within a 20 min driving period. Average concentrations exceeded 1000 ppm, so it is recommended to improve ventilation promptly while driving. (2) The MAPE of ARIMA and LSTM prediction results are 0.46% and 0.56%, respectively. The RMSE results are 19.62 ppm and 22.76 ppm, respectively. The prediction results demonstrate that both models effectively forecast changes in a vehicle’s interior environment CO₂, but the prediction accuracy of ARIMA is better than that of LSTM. The research findings provide theoretical guidance to traffic safety managers in selecting suitable models for predicting in-vehicle CO₂ concentrations and establish an effective in-vehicle ventilation warning control system.

Anti-money laundering (AML) remains a critical challenge in cryptocurrency ecosystems, where blockchain’s transparency paradoxically coexists with pseudonymity. Traditional methods often fall short in modeling the temporal and structural complexity of transaction networks. This paper introduces ChronoWave-GNN, a graph neural framework designed from the theoretical perspective of time-frequency representation learning. By combining wavelet-based frequency decomposition with temporal encoding, our model captures nonstationary and multi-scale patterns inherent in illicit financial activity. This dual-domain perspective enhances the expressive capacity of graph representations without relying on modular patching. We validate our approach on the Elliptic dataset, where ChronoWave-GNN achieves a test accuracy of 0.9802 and F1-score of 0.9799, surpassing prior state-of-the-art results. These findings suggest that unifying temporal dynamics and spectral compression offers a principled and effective pathway for robust AML in decentralized financial systems.

The increasing complexity of interconnected systems in the Internet of Things (IoT) demands advanced methodologies for real-time security and management. This study presents MAS-LSTM, an anomaly-detection framework that combines multi-agent systems (MASs) with long short-term memory (LSTM) networks. By training agents on IoT traffic datasets (NF-ToN-IoT, NF-BoT-IoT, and their V2 versions), MAS-LSTM offers scalable, decentralized anomaly detection. The LSTM networks capture temporal dependencies, enhancing anomaly detection in time-series data. This framework overcomes key limitations of existing methods, such as scalability in heterogeneous traffic and computational efficiency in resource-constrained IIoT environments. Additionally, it leverages graph signal processing for adaptive and modular detection across diverse IoT scenarios. Experimental results demonstrate its effectiveness, achieving F1 scores of 0.9861 and 0.8413 on NF-BoT-IoT and NF-ToN-IoT, respectively. For V2 versions, MAS-LSTM achieves F1 scores of 0.9965 and 0.9678. These results highlight its robustness in handling large-scale IIoT traffic. Despite challenges in real-world deployment, such as adversarial attacks and communication overhead, future research could focus on self-supervised learning and lightweight architectures for resource-constrained environments.

The widespread use of the Internet and social media has posed significant challenges to automated sentiment analysis, particularly in relation to detecting sarcasm in user-generated content. Sarcasm often expresses negative emotions through seemingly positive or exaggerated language, making its detection a complex task in natural language processing. To address this issue, the present study proposes a novel sarcasm detection model that combines bidirectional encoder representations from transformers (BERT) with gated graph neural networks (GGNN), further enhanced by a self-attention mechanism to more effectively capture ironic cues. BERT is utilized to extract deep contextual information from the text, while GGNN is employed to learn global semantic structures by incorporating dependency and emotion graphs. Experiments were conducted on two benchmark sarcasm detection datasets, namely Headlines and Riloff. The experimental results demonstrate that the proposed BERT-GGNN model achieves an accuracy of 92.00% and an F1 score of 91.51% on the Headlines dataset, as well as an accuracy of 86.49% and an F1 score of 86.59% on the Riloff dataset, significantly outperforming the conventional BERT-GCN models. The results of ablation studies further corroborate the efficacy of integrating GGNN, particularly for handling complex ironic expressions frequently encountered in social media contexts.