Corporate Social Responsibility (CSR) reporting has become an indispensable mechanism for organizations to communicate their sustainability initiatives. However, the growing volume and complexity of these reports necessitates the integration of Natural Language Processing (NLP)-driven text mining techniques to enhance transparency, comparability, and strategic decision-making. This study employs NLP and text mining methodologies to systematically analyze CSR reports with emphasis on environmental sustainability from major public listed companies in Taiwan. Utilizing Principal Component Analysis (PCA), this study classifies sustainability-related topics, extracts key Sustainable Development Goals (SDG)-aligned terms, and evaluates the textual similarities between CSR reports and SDG targets. Five SDGs encompassing 39 specific targets form the analytical framework, and 225 feature words are identified through text mining. The findings indicate that (1) automated CSR topic classification in Chinese is viable, though expert validation remains crucial for linguistic accuracy and semantic integrity; (2) SDG feature word distribution follows the Pareto Principle, with 10% of words contributing to 50.4%, and 28.4% accounting for 80% of total TF-IDF weights; (3) CSR reporting varies by industry, with financial holdings emphasizing sustainable management, energy supply, and water efficiency, while the electronics sector prioritizes waste reduction, recycling, and product lifecycle management; (4) PCA-based classification effectively aligns CSR reports with SDG targets, with textual similarity analysis proving more accurate than principal component scores. From a strategic business perspective, these findings offer critical insights for corporate leaders seeking to refine their sustainability strategies. NLP-based CSR analysis enables companies to benchmark their Environmental, Social, and Governance (ESG) performance against industry peers, identify sustainability gaps, and align corporate strategies with evolving regulatory landscapes and stakeholder expectations. Financial institutions can leverage these insights to develop sustainable finance mechanisms, while manufacturers can enhance circular economy practices by optimizing resource efficiency and waste management. Moreover, policymakers and investors can utilize NLP-driven text mining techniques to assess corporate sustainability efforts more systematically, ensuring greater accountability and fostering data-driven decision-making in ESG governance.

The popularity of social networking sites (SNSs) has increased rapidly. SNSs are a key part of daily life for many people around the world. The use of SNSs is already a global phenomenon. Drawing on social capital theory, this research empirically explored how structural capital (social network ties), relational capital (trust in SNSs, trust in members of SNSs, social identification, and social norms), and cognitive capital (shared language and shared goals) influence stickiness, which in turn affects addiction. This study introduced key moderators, privacy concerns and perceived security, to the relationship between stickiness and addiction. The authors empirically evaluated the proposed model by using survey data collected from SNS users. Structural equation modeling was applied to test the model. This study can provide a deeper understanding of SNS users' addiction behavior by focusing on social capital theory, privacy concerns, and perceived security and therefore contribute to both research and practice.

Mathematical morphology intends to extract object features such as geometric and topological structures in digital images. Given a set of target images and original images, it is cumbersome and time-consuming to determine the suitable morphological operations and structuring elements. In this paper, we propose deep morphological neural networks, which include a nonlinear feature extraction layer to learn the structuring element correctly and an adaptive layer to select appropriate morphological operations automatically. We demonstrate the applications of object recognition, including hand-written digits, geometric shapes, traffic signs, and brain tumor. Experimental results show the higher computational efficiency and higher accuracy of our developed model as compared against existing convolutional neural network models.

Adversarial attacks can fool convolutional networks and make the systems vulnerable to fraud and deception. How to defend against malicious attacks is a critical challenge in practice. Adversarial attacks are often conducted by adding tiny perturbations on images to cause network misclassification. Noise reduction can defend the attacks; however, it is not suited for all the cases. Considering that different models have different tolerance abilities on adversarial attacks, we develop a novel detecting module to remove noise by adaptive process and detect adversarial attacks without modifying the models. Experimental results show that by comparing the classification results on adversarial samples of MNIST and two subclasses of ImageNet datasets, our models can successfully remove most of the noise and obtain detection accuracies of 97.71% and 92.96%, respectively. Furthermore, our adaptive module can be assembled into different networks to achieve detection accuracies of 70.83% and 71.96%, respectively, on the white-box adversarial attacks of ResNet18 and SCD01MLP images. The best accuracy of 62.5% is obtained for both networks when dealing with the black-box attacks.

This research developed an effective supply chain management (SCM) operation model and a corresponding diagnostic methodology in the global competitive environment by combining three phases and methodologies. In Phases 1, a list of impact factors was collected from relevant studies, and a hierarchy for the current complex research issue was established. In Phase 2, an expert survey was conducted for enhancing the content effectiveness of the model. Subsequently, an analytic hierarchy process (AHP) was applied to determine the weights of those factors. In Phase 3, a self-assessment framework was developed to examine the effectiveness of the global SCM. Finally, a case was analyzed to verify the effectiveness of the self-assessment scales. The results show that the concept of global SCM tends to be broader and more comprehensive than traditional one. Especially in the coordination and cooperation between market, organization itself and suppliers, while using information technology as a communication tool.

This article describes how in recent years, enterprises have increasingly adopted social technologies to support customer relationship management (CRM) practices. To increase CRM performance, enterprises have had to develop appropriate social CRM strategies in emerging social network applications. This article investigates the alignment between social network applications and the “4C strategy” as a social CRM strategy in organizations. It intends to understand how alignment influences CRM performance. A total of 225 Taiwanese companies, which have adopted Facebook to interact with customers, were surveyed. According to the results, alignment between the 4C strategy and social network applications has a positive and significant impact on CRM performance. The results also suggest that organizations require a high-level integration of social network applications with the 4C strategy to achieve high CRM performance.

The traffic collision is one of the major people killers in the world. It is very helpful to prevent the traffic collision if a driver can identify the potential hazard and predict what will happen. This study presents a path estimation system which can analyze the current condition as well as predict the subsequent trajectory using the trained database. The repetitive and chronological characteristic of vehicle trajectories is utilized to learn the vehicle behaviors through clustering process. The framework consists of three key modules: trajectory extraction, vehicle behavior learning, and vehicle behavior prediction. Trajectory extraction applies vehicle detection and tracking to obtain the vehicle trajectories. Vehicle behaviors are learned and analyzed through a clustering process. Finally, through the multi-predictor model, vehicle behavior prediction predicts the vehicle positions in the following subsequence. Experimental results demonstrate that our proposed model can get the predicted trajectory with good accuracy.

This paper proposes an object verification method by using sparse representation (SR) which has been applied for object representation and recognition. However, SR dictionary does not show sufficient compactness. Our method comprises three major modules. First, we train the sparse matrix by using boost K-Singular Value Decomposition (boost K-SVD) to obtain a sparse vector set. Second, we combine two training sparse vector sets of the same and different objects from two views to generate a positive/negative combined sparse vector set. Finally, a Support Vector Machine (SVM) classifier is applied for verification. Our contributions are (1) obtaining a sparser vector set using K-SVD, (2) demonstrating the SR matrix with better Restricted Isometry Property (RIP), and (3) applying the SR matrix to the object verification process with high accuracy. The experimental results prove that our method has higher accuracy than the other methods.

This paper introduces a ribbon-based motion analysis approach to describe human body movements. Here, we assume that there are no markers on human body. We develop a system to extract the moving ribbons (extremities) by processing the difference between current image frame and reference image frame. By analyzing the moving ribbons on the key frames, we may produce the motion parameter curves for each joint on the ribbon. These curves may not be continuous due to ribbon-torso occlusion, and both the interpolation and extrapolation processes can be used to predict the missing parts.