This study aimed to explore the indirect effects of appearance-related anxiety (ARA) on Instagram addiction (IA) through sequential mediators, namely social media activity intensity (SMAI) and Instagram feed dependency (IFD). The study also aimed to provide theoretical explanations for the observed relationships and contribute to the understanding of the complex interplay between appearance-related concerns, social media usage, and addictive behaviors in the context of IA. A sample of 306 participants was used for the analysis. The results of the sequential mediation analysis (SMA) revealed several important findings. Firstly, the mediation model demonstrated that SMAI mediated the relationship between ARA and IA. However, there was no direct relationship observed between ARA and SMAI. Secondly, the analysis showed that IFD acted as a second mediator in the relationship between ARA and IA. Both ARA and SMAI had significant direct effects on IA, indicating their individual contributions to addictive behaviors. Furthermore, the total effect model confirmed a positive relationship between ARA and IA. This finding suggests that ARA has a direct influence on the development of IA. The examination of indirect effects revealed that ARA indirectly influenced IA through the sequential mediators of SMAI, IFD, and ultimately IA itself. The completely standardized indirect effect of ARA on IA through these mediators was found to be significant. Overall, this study provides evidence for the indirect effects of ARA on IA and highlights the mediating roles of SMAI and IFD. These findings contribute to our understanding of the psychological mechanisms underlying the complex relationship between appearance-related concerns, social media usage, and the development of IA.

Under the developing trend of artificial intelligence (AI) technology gradually penetrating all aspects of society, the traditional language education industry is also greatly affected ^[1]. AI technology has had a positive impact on college English teaching, but it also presents challenges and negative impacts. On the positive side, AI technology can provide personalized learning experiences, real-time feedback, and autonomous learning opportunities, and so on. However, it may also lead to a lack of communication between students and humans, resulting in a decline in students’ interpersonal skills, and cause students’ dependence on online learning resources as well as possible risks to student data privacy and security, and other negative impacts. To address these challenges, teachers can adopt the following countermeasures: improving teachers’ skills in the use of AI technology incorporated in the classroom, offering personalized instruction to reduce students’ dependence on AI technologies, emphasizing the cultivation of students’ humanistic literacy and interpersonal communication ability. Additionally, colleges and technology providers should strengthen data security and privacy protection to ensure the safety and confidentiality of student data. By implementing comprehensive measures, we can maximize the advantages of AI technology in college English teaching while overcoming potential issues and challenges.

Noise pollution in construction sites is a significant concern, impacting worker health, safety, communication, and productivity. The current study aims to assess the paramount consequences of ambient noise pollution on construction activities and workers’ productivity in Peshawar, Pakistan. Noise measurements have been recorded at four different construction sites in Peshawar at different times of the day. Statistical analysis and Relative Importance Index (RII) are employed to evaluate the data Risk variables, such as equipment maintenance, noise control, increased workload, material handling challenges, quality control issues, and client satisfaction. The results indicated that noise levels often exceeded permissible limits, particularly in the afternoon, posing significant worker risks. In addition, RII analysis identified communication difficulties, safety hazards, and decreased productivity as significant issues. The results show that noise pollution is directly linked with safety risks, decreased performance, and client dissatisfaction and needs immediate attention by authorities. This paper proposes a strategic policy framework, recommending uniform hand signals and visual communication methods without noise for workers, worker training about safety, and using wearable devices in noisy settings. Communication training for teams and crane operators, proactive quality control, and customer-oriented project schedules are also proposed. These recommendations aim to mitigate the adverse effects of noise pollution, enhance construction industry resilience, and improve overall operational efficiency, worker safety, and client satisfaction in the construction sector of Peshawar, aligning with policy and sustainable development objectives.

This study conducts a comparative analysis of various machine learning and deep learning models for predicting order quantities in supply chain tiers. The models employed include XGBoost, Random Forest, CNN-BiLSTM, Linear Regression, Support Vector Regression (SVR), K-Nearest Neighbors (KNN), Multi-Layer Perceptron (MLP), Recurrent Neural Network (RNN), Bidirectional LSTM (BiLSTM), Bidirectional GRU (BiGRU), Conv1D-BiLSTM, Attention-LSTM, Transformer, and LSTM-CNN hybrid models. Experimental results show that the XGBoost, Random Forest, CNN-BiLSTM, and MLP models exhibit superior predictive performance. In particular, the XGBoost model demonstrates the best results across all performance metrics, attributed to its effective learning of complex data patterns and variable interactions. Although the KNN model also shows perfect predictions with zero error values, this indicates a need for further review of data processing procedures or model validation methods. Conversely, the BiLSTM, BiGRU, and Transformer models exhibit relatively lower performance. Models with moderate performance include Linear Regression, RNN, Conv1D-BiLSTM, Attention-LSTM, and the LSTM-CNN hybrid model, all displaying relatively higher errors and lower coefficients of determination (R²). As a result, tree-based models (XGBoost, Random Forest) and certain deep learning models like CNN-BiLSTM are found to be effective for predicting order quantities in supply chain tiers. In contrast, RNN-based models (BiLSTM, BiGRU) and the Transformer show relatively lower predictive power. Based on these results, we suggest that tree-based models and CNN-based deep learning models should be prioritized when selecting predictive models in practical applications.

This paper aims to explore the relationship between corporate overinvestment and management incentives, focusing particularly on the influence of different ownership structures. Utilizing agency theory and ownership structure theory, this study constructs a theoretical framework and posits hypotheses on how management incentives might influence corporate overinvestment behaviors under different ownership structures. Listed companies from 2010 to 2020 were selected as the research sample, and the hypotheses were empirically tested using descriptive statistics, correlation analysis, and regression analysis. The findings suggest that a relatively concentrated ownership structure may encourage management to adopt more cautious investment strategies, thus reducing overinvestment behaviors; while under a dispersed ownership structure, the relationship between management incentives and overinvestment is more complex. This study provides new evidence on how management incentive mechanisms influence corporate decision-making in different ownership environments, offering significant theoretical and practical implications for improving internal control and incentive mechanisms.