Der-Fa Chen

The stability and reliability of temperature sensors in industrial cooling systems are critical to process quality, energy efficiency, and operational safety. However, existing approaches lack systematic stability metrics and intelligent predictive capabilities. This study proposes an AI-driven framework integrating stability feature engineering with machine learning models for fault identification and early prediction of temperature sensors in power plant cooling systems. The framework introduces three physics-based stability indicators—rolling standard deviation (σ_roll), variation intensity index (VII), and short-term variation magnitude (ΔT_short)—to quantify sensor signal quality. These features, combined with operational parameters, are used to train support vector machine (SVM) and Long Short-Term Memory (LSTM) models for binary classification. The framework is validated using over 260,000 one-minute records per unit collected from three parallel steam-turbine generating units (Units 1, 2, and 3) of the same coastal thermal power plant. Each unit is served by an independent once-through seawater cooling loop instrumented with redundant Pt-100 temperature sensors at the inlet and outlet manifolds; the three units differ in their operating profile—Unit 1 operates under variable load with frequent cold-start events, Unit 2 under moderate variable load, and Unit 3 under stable high-load conditions—with data collected at 1 min intervals from January to June 2025. Under an explicitly anomaly-positive evaluation, with the full confusion matrix reported for every unit and model, classification performance is limited and strongly unit-dependent. In real-time identification, AUC-based ranking ability varies across units (SVM AUC = 0.65, 0.75, and 0.98 for Units 1–3; LSTM AUC = 0.66, 0.31, and 0.52), but under the extreme class imbalance (anomaly rate ≈ 0.07–0.13% in the test partitions), the calibrated operating-point precision and F1-scores remain low for all unit–model combinations (F1 ≤ 0.26, MCC ≤ 0.28). McNemar’s test indicates statistically significant paired differences for Units 1 and 2 but not for Unit 3. These results show that, on this dataset, neither model attains reliable anomaly classification, and that all reported metrics must be interpreted together with the disclosed confusion-matrix counts and severe class imbalance. The primary contribution of the framework is therefore methodological—physics-based stability indicators, redundant sensor cross-checking, and an operational false-alarm analysis—rather than high-accuracy prediction, and the study highlights the difficulty of learning-based prediction for rare, rule-defined thermal sensor anomalies.

This paper proposes the Continuous Operational Feedback Loop (COFL) architecture, a fully localized, event-driven operational monitoring and response system for Direct-to-Consumer (D2C) commerce. The architecture integrates the n8n workflow engine with on-premise large language model (LLM) inference via the Ollama framework, forming a containerized stack deployable on commodity CPU-only edge hardware (~USD 1640). Using a multi-source dataset of 1800 records constructed from publicly available e-commerce corpora and evaluated with a silver-standard automated labeling protocol, empirical validation demonstrates an end-to-end latency of 3.22 s and a macro-F1 sentiment classification score of 0.836—representing 98.2% of the full-precision baseline and 94.0% of cloud GPT-4o API generation quality measured by ROUGE-L—at approximately 1/200th of the per-request inference cost. A systematic quantization ablation study across six model-quantization configurations establishes LLaMA 3 8B Q4_K_M as the Pareto-optimal selection for the target hardware. An Analytic Hierarchy Process (AHP) multi-criteria framework with criterion weights derived from published literature confirms the COFL implementation achieves a higher composite score than cloud API deployment under the stated evaluation assumptions. Failure mode and effects analysis (FMEA) is summarized to characterize system reliability under identified failure scenarios.

The pervasive integration of artificial intelligence across e-commerce ecosystems has fundamentally transformed the competitive landscape, rendering systematic and reproducible platform evaluation frameworks an operational necessity rather than an academic exercise. Conventional multi-criteria decision analysis approaches for e-commerce evaluation remain structurally constrained by their dependency on human expert panels, which introduce recruitment costs, cognitive biases, limited reproducibility, and the practical infeasibility of assembling genuinely multidisciplinary panels spanning e-commerce strategy, machine learning engineering, and financial technology simultaneously. This study proposes a novel decision support framework that integrates Large Language Model (LLM) multi-agent simulation with the Best–Worst Method (BWM) to derive reproducible priority weights for AI-driven e-commerce platform evaluation within a rigorous business intelligence architecture. Twelve domain-differentiated LLM agents—organized into three expertise groups representing e-commerce management, AI and machine learning technology, and digital payment systems—were instantiated with structured system prompts encoding professional domain knowledge and deployed across three independent simulation rounds to perform BWM pairwise comparisons across a comprehensive six-dimensional, 30-sub-criterion evaluation hierarchy. Inter-agent consensus was synthesized through geometric mean aggregation, with consistency verification conducted via BWM’s xi* indicator and inter-round stability assessed through coefficient of variation analysis. Results reveal that Transaction Security and Trust achieves the highest dimension-level weight (w = 0.248), followed by AI Recommendation Effectiveness (w = 0.213), with Personal Data Protection (G = 0.0750), Recommendation Accuracy (G = 0.0607), and Transaction Transparency (G = 0.0549) emerging as the three highest globally ranked sub-criteria. The aggregated consistency indicator xi* = 0.062 confirms logical coherence of the multi-agent judgment consensus, and all dimension weights exhibit CV values below 2.8%, demonstrating exceptional inter-round stability. Spearman rank correlations among the three domain-expertise groups exceed 0.92, confirming strong inter-group convergence. Sensitivity analysis under perturbations of ±10% and ±20% demonstrates that the top-five priority indicators are structurally stable. This study establishes LLM multi-agent BWM simulation as a methodologically rigorous, institutionally accessible, and computationally reproducible alternative to traditional expert elicitation for complex platform evaluation tasks.

Objectives: Smartphone addiction has become a salient mental health concern among university students. Although the Big Five personality traits are associated with problematic smartphone use, less is known about the psychological mechanisms linking personality to addictive smartphone behavior. This study examined whether mindfulness functions as a self-regulatory mechanism linking personality traits to smartphone addiction. Method: A cross-sectional survey was administered to Taiwanese university students (N = 665). Partial least squares structural equation modeling with bootstrapping was used to test direct and indirect associations among the Big Five traits, mindfulness, and smartphone addiction. Results: Mindfulness was negatively associated with smartphone addiction (β = −0.305, p < 0.001). Neuroticism showed a positive direct association with smartphone addiction (β = 0.278, p < 0.001) and a negative association with mindfulness (β = −0.372, p < 0.001), yielding a significant indirect effect via mindfulness (β = 0.113, p < 0.001). Openness also showed a positive direct association with smartphone addiction (β = 0.163, p = 0.003). Conscientiousness was positively associated with mindfulness (β = 0.224, p < 0.001). It exerted a significant indirect effect on smartphone addiction through mindfulness (β = −0.068, p < 0.001), while its direct path to smartphone addiction was not supported. Conclusions: Mindfulness appears to be a key self-regulatory mechanism linking personality traits to smartphone addiction. These findings support the practical value of mindfulness-based approaches for promoting healthier digital habits and improving student digital well-being.

This study aims to develop a preliminary list of indicators suitable for green restaurants in Taiwan. The research methodology includes expert consensus (Delphi method) and incorporates interviews with field experts. An analysis of the responses provided by these industry experts led to the identification of five dimensions of evaluation indicators for green restaurants. The K–S test involves using a z-test on ordinal variables for single samples to determine whether the sample distribution diverges from the frequency distribution. This study analyzed the responses provided by the interviewed experts to identify and extract evaluation indicators for green restaurants. The extracted indicators comprise five dimensions (resource management, ingredient and product selection, environmental and indoor quality, green certification and management, and customer awareness and participation), 15 sub-dimensions, and 70 detailed indicators. The research results can serve as a reference for course planning in related programs at universities and colleges, as well as for industry planning of green restaurants, and as a reference for the promotion of national sustainable environmental policies in Taiwan. Therefore, based on the results of this study, recommendations are provided for educational institutions related to green restaurants, official organizations related to green restaurants, the industry related to green restaurants, and future researchers.

In recent decades, scholarly interest has grown in the psychological components of happiness. Savoring belief, or the capacity to attend to, appreciate, and enhance the positive experiences in one’s life, along with resilience and meaning in life, have emerged as significant predictors of enhanced happiness among diverse populations. This research examined the interrelationships among savoring belief, resilience, meaning in life, and happiness. A sample of 561 students from 75 universities in Taiwan, comprising 361 female and 200 male participants with an average age of 20.88 years, participated in an online survey. The study employed various instruments, including the Savoring Belief Inventory, the Subjective Happiness Scale, the Brief Resilience Scale, and the Meaning in Life Questionnaire. These instruments were translated into Traditional Chinese using a back-translation method and subsequently validated for accuracy by specialists in the field. Analysis of the data using Hayes’ PROCESS Model 6 revealed several key insights: (1) savoring belief positively influenced happiness, resilience, and meaning in life with resilience further enhancing happiness and meaning in life; (2) resilience served as a significant mediator in the relationship between savoring belief and happiness; (3) meaning in life significantly mediated the relationship between savoring belief and happiness; (4) a sequential mediation model illustrated the mediating effects of resilience and meaning in life on the relationship between savoring belief and happiness. This study illustrates that, much like a garden requires water, sunlight, and care to flourish, our happiness is cultivated through enhancing our ability to savor the good moments, rebound from challenges, and find deep significance in our lives. We can significantly boost well-being by fostering these qualities—savoring belief, resilience, and a sense of meaning. These findings are particularly relevant for educators, highlighting the critical need to develop these traits in students to promote greater happiness and fulfillment in their lives. Discussions included theoretical implications, educational implications, and avenues for future research.

This research investigates the control of thickness and weight in plastic extrusion vacuum-thermoforming products to identify optimal key parameters for cost reduction and energy savings. The initial step involves identifying crucial influencing factors. In this step, the Delphi technique was employed through a questionnaire administered to a panel of expert scholars to ensure minimal error and maximal reliability in determining key influencing factors. Consensus was sought to establish appropriateness and consistency. Subsequently, the Taguchi method was applied for quality design and planning of the extrusion vacuum-forming process. The experimental design parameters were selected using an L₁₈ (2¹ × 3⁷) orthogonal array, and the desired quality characteristics were determined. Comparative analysis of quantitative production data from two consecutive experiments was conducted, and based on F-values and contribution analysis, the combination of control factors maximizing the Signal-to-Noise (S/N) ratio was identified. The objective is to seek optimal parameters for improving the quality of the plastic polypropylene (PP cup lid) manufacturing process, reducing process variability, and identifying the most robust production conditions. Through multiple actual production prediction experiments, it was determined that five control factors, “polypropylene new material ratio,” “T-die lips adjustment thickness”, “mirror wheel temperature stability”, “molding vacuum pressure time”, and “forming mold area design”, contribute to the maximization of the S/N ratio, i.e., minimizing variability. Statistical validation confirms a significant improvement in product quality and weight control. Noteworthily, the quality control model and experimental design parameters established in this study are also applicable to other plastic products and bio-based materials, such as PET, HIPS, and biodegradable PLA lids with added calcium carbonate. The results of the experimental production demonstrate its ability to consistently control product weight within the range of 3.4 ± 0.1 g, approaching the specified tolerance limits. This capability results in approximately 2.6% cost savings in product weight, contributing significantly to achieving a company’s KPI goals for environmental conservation, energy efficiency, and operational cost reduction. Therefore, the findings of this study represent a substantial and tangible contribution.

A circular economy is a concept of sustainable development that enhances a company’s image and influences consumers’ attitudes toward the corporation. Trust, commitment, and identification in relationship marketing are key factors for fostering long-term relationships. This study aimed to examine the relationship quality and attitudes between higher education students’ personal environmental perspectives and circular economy products, further investigating their impact on the purchase intention of circular economy products. Therefore, this research proposed and tested a theoretical framework that combined variables, including environmental concern, environmental awareness, relationship quality with circular economy products, and attitudes, to determine the influencing factors on higher education students’ purchase intention of circular economy products. This study adopted a survey design method and employed partial least squares structural equation modeling (PLS-SEM) to analyze data from 443 higher education students from Taiwan and Thailand. The results indicated that environmental awareness and environmental concern are critical antecedent factors for attitude and relationship quality (trust and identification) in circular economy products. In addition, trust and attitude in circular economy products have a significant impact on commitment. Furthermore, trust, commitment, and identification with circular economy products are crucial determining factors for the purchase intention of circular economy products. The findings contribute to a deeper understanding of the factors influencing higher education students’ purchase intention of circular economy products, thereby achieving the goal of environmental sustainability.

The booming development of livestream e-commerce has attracted considerable academic attention, but research on how social presence affects consumers’ purchase decisions is limited. To this end, this article proposed a theoretical framework for the influence of social presence on consumers’ purchasing decisions based on social presence theory. In this study, structural equation modeling was carried out on 390 data collected from a questionnaire to verify the mechanisms by which social presence influences purchase intention. The study found that social presence enhances consumer identification, which in turn enhances consumers’ purchase intention. Meanwhile, this study not only verified the mediating role of consumer identification but also tested the positive moderating role of self-improvement by livestream e-commerce. This study found that a live streaming host’s social presence influences the consumers’ purchase behavior, and it provides decision support for companies to sell products through livestreaming platforms, which has important theoretical significance and practical application value.

This study took food-grade polypropylene packaging products as the research project and discussed how to control the polypropylene extrusion sheet thickness and vacuum thermoforming quality and weight. The research objective was to find the key factors for reducing costs and energy consumption. The key aspects that may influence the polypropylene extrusion molding quality control were analyzed using literature and in-depth interviews with scholars and experts. These four main aspects are (1) key factors of polypropylene extrusion sheet production, (2) key factors of the extrusion line design, (3) key factors of polypropylene forming and mold manufacturing, and (4) key factors of mold and thermoforming line equipment design. These were revised and complemented by the scholar and expert group. There are 49 subitems for discussion. Thirteen scholars and experts were invited to use qualitative and quantitative research methods. A Delphi questionnaire survey team was organized to perform three Delphi questionnaire interviews. The statistical analyses of encoded data such as the mean (M), mode (Mo), and standard deviation (SD) of various survey options were calculated. Seeking a more cautious research theory and result, the K-S simple sample test was used to review the fitness and consistency of the scholars’ and experts’ opinions on key subitem factors. There are ten key factors in the production quality, including “A. Main screw pressure”, “B. Polymer temperature”, “C. T-die lips adjustment thickness”, “D. Cooling rolls pressing stability”, “E. Cooling rolls temperature stability”, “F. Extruder main screw geometric design”, “G. Heating controller is stable”, “H. Thermostatic control”, “I. Vacuum pressure”, and “J. Mold forming area design”. The key factors are not just applicable to classical polypropylene extrusion sheet and thermoforming production but also to related process of extrusion and thermoforming techniques in expanded polypropylene (EPP) sheets and polylactic acid (PLA). This study aims to provide a key technical reference for enterprises to improve quality to enhance the competitiveness of products, reduce production costs, and achieve sustainable development, energy savings, and carbon reductions.