34254 research outputs found
Sort by
Discrepancies in Expectations and Actual Experiences on Dormitory Services Quality of University on-Campus Residents
No full text\ue3\ue3To provide students living on campus with the best living experiences, schools often conduct satisfaction surveys to understand how to enhance students\ue2 living environments under limited resources. This work takes a previous housing survey of an unnamed Taiwanese university as an example. Though the survey results did show a decline in negative views, most of the criteria are not as objective. This study aims to objectively evaluate the over all improvement of dormitory services to understand the discrepancies in students\ue2 expectations and actual experiences. The inconsistencies can serve as a reference for future ameliorations.
\ue3\ue3Online quantitative questionnaires, created based on the SERVQUAL model, were adopted to assess how much residents value and aresatisfied with dormitory facilities and services. A total of 485 responded surveys were received. According to Cronbach's alpha, responses showed good internal consistency. The findings suggest that the dormitories provide sufficient cookingareas, accessible environments, and amenities.
\ue3\ue3Study results can be referenced for administrative staff when assigning rooms and coordinating events in the future. Further, dormitory staff can prioritize categories that had thehighest discrepancies to optimize and streamline management and services within dormitories
Study on the Application of Biometric Technology to Administrative Investigations on the Border\uef\ubcBased on the Protection of Basic Human Rights
No full text\ue3\ue3With the advancement of technology, biometric identification technology has been widely applied in various fields as a supplementary method for confirming personal identity when entering or leaving a country. It has become an international mainstream trend. However, the collection, use, and processing of these personal privacy information have also raised issues concerning human rights protection and information security.
\ue3\ue3This study aims to explore how biometric identification technology is applied to administrative investigations on the border, and how it interferes with basic human rights of individuals. It also examines the constitutional basis for such interference with basic human rights. By clarifying the scope of exercise of administrative investigation powers on the border and implementing effective legal regulations to safeguard the basic human rights of individuals, this study seeks to provide specific recommendations and measures to address the administrative investigation practices and information security issues related to the collection, use, and processing of individuals' biometric data by our country's border law enforcement agencies. The goal is to minimize potential impacts, assist in smooth policy implementation, enhance the quality and efficiency of border clearance services, strengthen border management, prevent transnational crimes, and ensure the protection of human rights, while striking a balance between national security maintenance and the protection of basic human rights of individuals
The Influence of Social Media Marketing Strategies on User\ue2s Brand Knowledge and Purchasing Intention.-Moderated by Sports Involvement.
No full textWith the advancement of Internet technology and the maturity of web technology, the changes in business behavior have been promoted, and electronic commerce has begun, which has changed the way of business-to-marketing. This study explores the relationship between brand knowledge and purchase intention for sporting goods brands to adopt new marketing methods such as KOL and social media marketing fan professional management.
The empirical results of this study show that KOL has no significant effect on improving brand awareness and brand image, while social media marketing has a positive and significant effect on improving brand awareness and brand image. In addition, both brand awareness and brand image have positive and significant effects on purchase intention. However, sports involvement cannot moderate the relationship between brand awareness, brand image and purchase intention.
This study can provide sports brands with different results when adopting new marketing strategies, especially when the two strategies of KOL and social media marketing are in parallel, and give sports brands a choice of marketing strategies. As long as brand awareness and brand image are strengthened, consumers' purchase intention can be enhanced
Semi-automated quantification of cerebral oxygen extraction fraction with susceptometry-based MR oximetry
No full textWhole-brain oxygen extraction fraction (OEF) is defined as the oxygen consumption fraction, which can reflect the balance between oxygen supply and metabolism. Therefore, OEF has been viewed as one of the physiologic biomarkers for assessing brain functions and oxygen metabolism. In this study, image data were collected by using a multi-echo gradient-echo sequence. The venous oxygen saturation level (SvO2) at the superior sagittal sinus (SSS) was measured via susceptometry-based oximetry (SBO), along with the arterial oxygen saturation level, yielding OEF. Specifically, SvO2 is associated with the phase difference between SSS and the surrounding area; however, manual segmentation of interest of regions (ROIs) might introduce additional bias into data analysis, implying that SvO2 and the subsequent OEF quantification are susceptible to operators. Hence, we developed a semi-automated analysis to mitigate operator-induced bias, which can improve the reliability of physiologic parameters. There were 20 healthy subjects enrolled in this study (28.9\uc2\ub15.4 years old). Data analysis was conducted by an experienced (operator 1) and an inexperienced (operator 2) operators utilizing manual and semi-automated methods. The results showed no significant difference in OEF between the two methods (manual vs. semi-auto: 36.5\uc2\ub110.1% vs. 36.2\uc2\ub110.1%, P = 0.29) performed by operator 1. On the other hand, OEF from operator 2 was significantly different (manual vs. semi-auto: 37.8\uc2\ub110.6% vs. 36.1\uc2\ub110.5%, P = 0.001). In addition, SvO2 was smaller from operator 2 than from operator 1 (Operator 1 vs. Operator 2: 62.3\uc2\ub19.9% vs. 60.9\uc2\ub110.4%, P = 0.001). Finally, there was no significant difference in OEF between the two operators when using the developed pipeline (36.2\uc2\ub110.1% vs. 36.1\uc2\ub110.5%, P = 0.49). The results suggest that operator-induced bias can be reduced with the semi-automated analysis
Design and Implementation of a Pet Adoption and Recovery Service System Using Data Visualization
No full textIn recent years, the number of people keeping pets has been steadily increasing. From 2012 to 2022, the total number of registered dogs and cats in Taiwan increased by over 1.9 million. However, along with the rise in pet ownership, the issue of pet abandonment has also emerged. These abandoned pets not only contribute to environmental clutter but can also pose dangers to pedestrians, causing problems to the local communities. Many organizations concerned about these issue have developed platforms for adopting and fostering stray animals, aiming to reduce the number of strays and provide more choices for potential pet adopters while helping these animals find better living conditions.
However, existing systems face limitations due to missing data and data uncertainties, restricting their functionality. In this work, a visualized pet adoption & recovery service was developed, displaying the uncertainty of animal ages and using maps to show the locations where the animals were found. Additionally, the system heavily relies on icons and colors instead of text to enhance data readability. This design allows users to view the data from different perspectives, making it easier and faster to find their desired pets. At the same time, it helps pet owners reunite with their missing pets quickly through the system. The research evaluated and validated the effectiveness and user-friendliness of our developed visual design and system
Design of Four-Axis Vibration Tools with Force Sensing Capability Based on Modal and Harmonic Analysis
No full textNowadays, the requirements for process precision are higher than before. Compared with traditional processing, tools with vibration functions can obtain better precision and surface roughness. In the era of smart manufacturing and the internet of things, a tool with the function of force sensing can instantly give feedback on the current processing situation and predict the tool's life to increase the advantages of automated manufacturing. Therefore, the goal of this research is to design a four-axis vibrating tool with force sensing function. The vibration mode includes two axial bending vibration modes, one axial longitudinal vibration mode, and one axial torsional vibration mode. Those four vibration modes lead the tool to produce elliptical vibration or specific axial vibration, which can be applied in different processing or meet different requirements. At the same time, to achieve force sensing capability, lead zirconate titanate (PZT) is used as the actuator and force sensor. PZT can drive the tool through the converse piezoelectric effect, and PZT can detect force through the direct piezoelectric effect.
In this study, the size of the tool is decided by the analytical solution and finite element method so that the vibrations in different modes can reach the same natural frequency and produce an elliptical vibration. The actual tool may have a different natural frequency compared with the result of the finite element method due to processing errors and assembly errors, so frequency adjustment structures are designed to adjust the natural frequency of the tool and reduce the effect of processing errors.
Finally, experiments were carried out to verify the performance of the designed tool. The vibration amplitude of the third mode of bending in the horizontal direction can reach up to 0.528 \uce\ubcm, and the vibration amplitude of the third mode of bending in the vertical direction can reach up to 0.710 \uce\ubcm. The maximum vibration amplitude of the first mode of longitudinal vibration in the Z-axis is 1.148 \uce\ubcm, and the Z-axis torsional vibration has a twist angle of 0.121 mrad. At the same time, the frequency adjustment structure can make the natural frequency of the third mode of the bending vibration in the horizontal direction have a 1 kHz adjustment range, and the Z-axis longitudinal vibration first mode vibration has an adjustment range of 0.8 kHz. In terms of force sensing capability, it can detect the force of three axes. When the PZT is driven at the anti-resonant frequency, the sensitivity can reach up to 0.2982 V/N
A Control of Octal-Switch Redundant Submodules to Tolerate an Open-Circuit Fault for Modular Multilevel Converters
No full textModular Multilevel Converter (MMC) is highly regarded for its stability, efficiency, modularity, and reliability in various fields. However, the risk of switch failures due to a high number of modules can cause system shutdown and grid damage.
This thesis proposes a fault-tolerant strategy using an octal-switch redundant submodule to handle Insulated Gate Bipolar Transistor (IGBT) open circuit faults. By detecting faults through residual analysis and compensating for voltage shortfalls, grid damage is prevented. The faulty submodule is bypassed using active fault diagnosis, and the redundant module takes over, achieving open-circuit fault tolerance in 1.5 cycles. The faulty submodule can then be hot-swapped, boosting MMC's fault tolerance and reliability in power systems.
The adopted system specifications are as follows: The line-to-line AC voltage is set at 380 volts, the DC voltage connected within each module is 400 volts, and each arm connects two modules. The DC terminal voltage is set at 800 volts, and the converter output power is 6 kW
Team Development in a Temporary Organization: The Case of the 2023 Chiayi Grasstraw Festival Collaborative Curatorial Team
No full textIn today's dynamic work landscape, collaborative efforts have become paramount, with professionals from diverse fields often uniting to address multifaceted challenges. This study centers on the curatorial team of the 15th Grasstraw Festival, a celebrated arts and cultural event in Chiayi. Originating as a performing arts hub, the festival, initiated by Our Theatre, has evolved by engaging multiple Chiayi-based organizations in co-curatorial roles, endorsing the benefits of inter-organizational collaboration.
However, the pandemic posed unforeseen challenges for the 15th edition of the festival. Assembled under stringent time pressures, the curatorial team comprised members from varied backgrounds, both new and seasoned. Such dynamics present potential challenges to team growth and cohesion. This research delves into the team\ue2s evolutionary stages through the lens of temporary organizations. Employing qualitative research methodologies, it discerns five pivotal stages: Preliminary Initiation, Content Planning, Curatorial Execution, Performance Linking, and Restoration & Reorganization.
Key insights reveal that when temporary teams are assembled swiftly with a blend of new and experienced members, there's a pronounced focus on task completion, often sidelining interpersonal cohesion. Moreover, for teams spanning multiple organizations with a vision for sustained collaboration, the Restoration & Reorganization phase is vital. This phase allows for recalibration, addressing inherent cross-collaborative ambiguities and setting the stage for future endeavors. Drawing from these insights, the research provides valuable theoretical perspectives and actionable implications and paves the way for future studies
On the Asymptotic Behavior of Allocation Scheme with Geometric Distributions
No full textConsider an allocation scheme of n indistinguishable balls independently dis-
tributed into N different boxes with equal probability. Then the contents of the boxes follow the multinomial distribution, which is a usual allocation scheme. A generalization of this scheme was introduced by Kolchin (See Kolchin [2] and Pavlov [3]), allowing the allocation probability to be unequal. We will focus on one of Kolchin\ue2s model with the allocation probability driven by geometric distributions. We will show the asymptotic behavior of the order statistics of the number of balls allocated to boxes
A Multi-agent Deep Reinforcement Learning with Adaptive Local Cooperation for Traffic Signal Control
No full textMany recent studies show that reinforcement learning (RL) is an effective approach to find a proper traffic signal control strategy. RL methods can obtain real-time environmental information and determine the control strategy based on those state changes. Conventional RL methods use centralized control, although global information is accessible, increasing the size of the environment will considerably raise the dimensions of the observation and action space. The RL agent faces the challenge of data failure and lost structure information of traffic network when gathering global state from a real-world environment. Recent research has attempted to overcome this limitation with deep reinforcement learning (DRL) and multi-agent reinforcement learning (MARL). These methods divide the environment into a few smaller regions and each local agent can only observe part of the environment. However, learning with partial information might cause local agents to make conflicting decisions with the neighborhood. The proposed method is based on a multi-agent reinforcement-learning algorithm with a communication protocol, and messages exchanged between local agents are selected by networks. The proposed method uses a graph attention network to retrieve the cooperation value based on present traffic conditions. With the dynamic cooperation value and neighborhood messages, local agents can pay more attention to important information and ignore noises. The proposed method further applies this cooperation value to reward calculation, encouraging local agents to learn cooperative decisions. This study compared with other RL-based methods on three different scales of datasets. The experimental results show that the proposed method can improve coordination among local agents and stabilize the learning process