34254 research outputs found
Sort by
Design and development of Wearable Point-of-Care Devices for Real-time Health Monitoring Utilizing AgNW based hybrid nanocomposites
No full textAbstract
The recent COVID-19 outbreak emphasized the need for personalised physiological monitoring using wearable and flexible electronics for health assessment and early detection of numerous diseases. There exist several imperative markers such as respiration, body temperature, respiratory associated motions and blood pressure among others, that encompasses a wealth of physiological information reflective of health and possible diseases. These markers can be exploited using point-of-care (POC) devices to enable real time health monitoring. Currently used POC devices demonstrate limited practicality for real time health monitoring due to several factors including susceptibility to virus contamination, wearing discomfort, intricate configuration and lumbering gadgets. Wearable smart sensors-based POC devices have great potential for real time health monitoring applications which facilitates early diagnosis of diseases thus saving lives. Consequently, comfortable and wearable smart sensor for real-time health monitoring are essential to empower the existing healthcare technologies. This dissertation demonstrates the design and development of non-invasive and highly sensitive wearable patch type sensors for POC health monitoring utilizing silver nanowires (AgNWs) based nanohybrid conductive inks. The integration of these nanohybrid conductive inks with different polymers are explored to design functional hybrid nanocomposites for real time POC healthcare applications, including body temperature assessment, respiratory monitoring and pulmonary function analysis. Here, Ag-Au core-sheath nanowire (Au@AgNWs) based hybrid ink is synthesized by cost effective conformal deposition to design a highly sensitive patchable and biocompatible temperature sensor. A novel temperature sensing mechanism based on dynamic internanowire distance(s) of the PEG coated Au@AgNWs percolation by means of capillarity force as a result of the glass transition temperature of thermosensitive PEG is demonstrated. Moreover, AgNWs-AgF and AgNWs-Mxenes based two individual nanohybrid inks are designed to fabricate ultrasensitive patchable strain sensors for real time breath monitoring and pulmonary function analysis utilizing the synergistic effect of 1D-2D nanohybrid. A novel methodology which translates observed one-dimensional strain into corresponding pulmonary volumes for the calculation of forced volume capacity (FVC), forced expiratory volume (FEV1) and peak expiratory flow (PEF) is proposed. Conclusively, a POC device for real time breath monitoring and pulmonary function analysis has been developed, which shows high correlation with commercial device. The designed POC device can also be implemented on lab rats in normal and anesthetized state for respiratory monitoring, demonstrating its practical feasibility for human and other species. This thesis aims to facilitate interdisciplinary research in cutting-edge nanohybrid materials for the advancement of real time healthcare monitoring systems
An automatic detection model for cardiovascular disease based on time-frequency domain features of PCG signals
No full textCardiovascular disease is one of the leading causes of death today. Therefore, early detection and treatment are essential to improve patient survival rates. The stethoscope is a common diagnostic tool that converts heart sounds collected through the stethoscope into a phonocardiogram (PCG). However, even experienced doctors can sometimes make misdiagnoses. Therefore, in the past few decades, it has been a popular research topic to develop automatic heart sound detection models by analyzing phonocardiograms and combining machine learning methods. Although some studies have shown good results, yet there are still shortcomings, such as most studies only consider clean data and conduct research under ideal conditions, which don't correspond to the clinical situation. In this study, we propose a new method that uses time-frequency domain analysis to extract multiple features and then combines a stacked recurrent neural network model and machine learning methods to improve the accuracy and practicality of heart sound recognition.
We apply the proposed methods to 838 clinical data; the results show that a deep learning model combined with XGboost obtains the best detection results, achieving an accuracy of 85.26% (sensitivity: 79.22%, specificity: 87.93%) on the test set. Compared with the results of experienced cardiologists, achieving an accuracy of 81.04% (sensitivity: 68.98%, specificity: 98.65%), our proposed method has higher accuracy and sensitivity. We apply the proposed approach to the famous 2016 PhysioNet/CinC Challenge public dataset to further validate the model's reliability. Deep learning combined with XGboost achieves an accuracy rate of 95% (sensitivity: 91.81%, specificity: 98.88%)
Intermedia Agenda setting of Issue ownership and Trait Ownership of Political Candidates: A case study of Taiwan's 2020 presidential election.
No full textIn the dynamic landscape of thriving online communities, diverse internet ecosystems have significantly influenced the lives of the general public. With the shift in media paradigms, the role of traditional media in agenda-setting and its network agenda-setting effects with social media remain a focal point of this study. This research combines the concept of issue ownership and trait ownership of presidential candidates in Taiwan in 2020 to investigate the cross-media network agenda-setting effects between traditional media and online forums.
This study employs a dictionary-based computer-assisted content analysis using web crawling programs to retrieve relevant articles from four traditional media sources, including "Liberty Times," "United Daily News," "Apple Daily," and "TVBS," which are associated with presidential candidates Tsai Ing-wen and Han Kuo-yu. Additionally, articles related to the two candidates are collected from the gossip board of the PTT forum, using a text analysis platform developed by the School of Management at National Sun Yat-sen University. The data collection period spans the three months leading up to the 2020 presidential election. The text data is segmented using the jieba package in the R language, followed by emotion analysis using the Chinese version of the Linguistic Inquiry and Word Count (CLIWC). Issues and traits owned by each candidate are identified using issue and trait dictionaries. Cross-lagged correlation analysis is then conducted to examine the cross-media network agenda-setting effects between the four traditional media outlets and the PTT forum.
This study uncovers the cross-media agenda-setting effects between mainstream news media and the online forum PTT. In terms of issue ownership networks, the findings reveal that news media exert a significant influence on PTT. Conversely, within the realm of trait ownership networks, during the T1-T2 timeframe, PTT influences news media, while during the T2-T3 timeframe, news media demonstrates a stronger influence on PTT. Moreover, irrespective of whether it is news media or PTT, the quantities of issue and trait ownership are consistently higher for Tsai Ing-wen compared to Han Kuo-yu
Spatiotemporal Variation and Speciation of Atmospheric Mercury Transport between Kaohsiung Harbor and Urban Areas
No full textKaohsiung City is the largest industrial city and international commercial harbor in Taiwan. The harbor's throughput ranks the 16th in the world, accounting for nearly 50% of the import and export cargo handling volume and over 54% of the container handling volume in Taiwan. The harbor is a major source of pollution, with emissions from nearby industrial areas, vehicles, ships, cargo handling, and transportation contributing to poor air quality in the harbor and nearby urban areas.
Our research team conducted atmospheric speciation mercury sampling from January to October 2022 to investigate the inter-transport between Kaohsiung Harbor and urban areas in different seasons. Four sampling sites included sites CH and ZH in the nearby harbor area, and sites MC and FS in the urban area, which allowed us to examine the pollution characteristics and potential sources of atmospheric mercury transported between the harbor and urban areas.
The seasonal variations of atmospheric mercury concentrations in the harbor and urban areas were ordered as: winter>autumn>spring>summer. The average concentrations of GEM, GOM, and PBM were 7.13\uc2\ub12.2 ng/m3, 331\uc2\ub1190 pg/m3, and 532\uc2\ub1301 pg/m3, respectively. The concentrations of atmospheric mercury were mainly influenced by prevailing winds, local sources, and atmospheric dispersion. In terms of spatial distribution, the concentrations of atmospheric mercury were ordered as: ZH>CH>FS>MC. GEM was the primary species of TAM accounting for around 85-94%, indicating significant anthropogenic emissions from the harbor areas.
Based on backward trajectory, pollution rose, and correlation analysis, we found that in winter and spring, polluted air masses primarily came from the northeast along the western Taiwan, possibly co-influenced by local sources and long-range transport. In summer, air pollutants originated from the southeast, likely influenced by coastal and industrial emissions. In fall, air pollutants came from offshore waters in western Taiwan, potentially influenced by emissions from southern China and southeastern coastal cities. In the urban areas, except for summer, air pollutants mainly originated from the west, indicating the influence of emissions from the harbor area. In summer, both industrial and harbor areas contributed to the air pollution in the urban area.
PMF and PCA analysis revealed that major sources in the harbor area were primarily attributed to ship traffics, cargo handling, automobile emissions, and nearby industrial complex. The main factors were coal-fired industrial boilers and mobile sources. In the urban area, major sources were influenced by emissions from the harbor area and local human activities. Mobile sources and sulfur-containing fuel and waste combustion were the main sources in the urban area. Site MC that is close to the harbor area exhibits similar sources as the harbor area, while site FS was mainly influenced by local emissions
A first-principles study of the effect of P, Cu, Cr solute atoms and vacancies on the screw dislocation of bcc Fe
No full textWe utilized first-principles calculations to investigate the interactions between dislocations and atoms (phosphorus, copper, chromium) as well as vacancies in ferrite. In electrical steel, material properties such as hardness, strength, resistivity, and iron loss are commonly enhanced through solid solution, which reduces the iron loss. However, solid solution atoms in iron can impede dislocation movement, leading to processing challenges in electrical steel production. Drawing upon our laboratory's previous work, we employed atomic-scale first-principles calculations to characterize the core region of dislocations, which cannot be explained by elastic theory alone. We established structures with varying dislocation densities and examined their effects. Our calculations revealed that the dislocation core induces an increase in the magnetic moment of the surrounding region, thereby strengthening the magnetic moment. Subsequently, we introduced common industrial alloying elements (phosphorus, copper, chromium) and vacancies into the dislocation structure through substitution to form a solid solution. We performed calculations to explore the interaction between solid solution and dislocations at different distances. Various properties were computed, including the solid solution formation energy, dislocation dipole energy, interaction energy between atoms and dislocations, and the magnetic moment of the structure. Regarding solid solution formation energy, we found that phosphorus, copper atoms, and vacancies exhibit a preference for the vicinity of the dislocation core, whereas chromium atoms are unaffected by the distance. Analysis of the dislocation dipole energy demonstrated that phosphorus, copper atoms, and vacancies contribute to stabilizing the dislocation structure, while chromium atoms have a minimal impact on dislocations. The interaction energy calculations revealed that phosphorus, copper atoms, and vacancies exhibit attractive forces towards dislocations when they are in solid solution in close proximity. This suggests that they exert a pinning force on dislocations. When the atoms are farther away from the dislocation core, the interaction between dislocations and these atoms decreases, and there is negligible interaction between chromium atoms and dislocations. Finally, the results of magnetic moment calculations show that vacancies and chromium atoms can enhance the magnetic moment in the surrounding region, while the strengthening effect of phosphorus and copper atoms is less pronounced. Since phosphorus, copper, and vacancies are non-magnetic, their solid solution in the structure leads to a decrease in the overall magnetic moment. However, due to the antiferromagnetic nature of chromium atoms, they cause a greater reduction in the magnetic moment. While solid solution leads to a decrease in the magnetic moment, the calculated magnetic flux density shows minimal differences between the solid solution and pure iron dislocations. Therefore, the impact on the magnetic properties is insignificant. Finally, chromium is found to be the most suitable element based on the required properties of the electrical steel, which include low pinning force, the ability to increase resistance, and minimal impact on magnetic properties
The Impact of COVID-19 on Volatility Trading Indicators Using Event Study
No full textThis study employs the event study method to investigate the impact of the COVID-
19 pandemic on trading volatility products. The empirical analysis focuses on VIX futures from 2017 to 2020 as the research subject. The World Health Organization's declaration of the COVID-19 outbreak on March 11, 2020, is considered the starting point of the pandemic phase. By utilizing the Trading Volatility index as a daily market entry and exit indicator, the study examines whether trading VIX futures during the COVID-19 period generates abnormal returns.
The empirical findings indicate that using the trading indicators as the basis for daily market entry and exit shows different effects before and during the pandemic, particularly when settling returns based on prices in the last 45 and 60 minutes of the trading day. Furthermore, to enhance the usability of the trading indicator, this study incorporates trading thresholds and swing trading into the trading strategy. The results demonstrate that the impact of the COVID-19 pandemic on trading VIX futures becomes more pronounced when these additional trading elements are included. Therefore, empirical evidence based on the Trading Volatility indicators confirm the influence of the COVID- 19 pandemic on trading VIX futures
3D Whole Brain Relaxometry Using Radial Triple Echo Steady State Scan
No full text\ue3\ue3Quantitative magnetic resonance imaging (QMRI) technology can quantitatively and objectively assess some changes in specific tissues that have different physiological characteristics or associated features based on images. Relaxometry is one of the QMRI techniques, which quantify relaxation times (e.g., T1 and T2). Quantitative maps will be estimated to examine tissue properties.
\ue3\ue3In clinical applications, different imaging protocols are implemented to acquire different image contrasts, but in general long scan time is required to obtain a complete information. In this study, we use TESS pulse sequence that acquires three pathway signals. TESS obtains T1 and T2 information in one single scan and we apply this sequence for brain imaging. In addition, in order to reduce the motion artifacts caused by cerebrospinal fluid flow, radial sampling scheme is used and combined with the probability density function (PDF), which determines the number of sampling points on each spoke to reduce the scan time.
\ue3\ue3To validate the our method, we make a number vials of gelatin phantom doped with different NiCl2 concentration. We have different sampling TESS scans and reference scans at 3.0T. Due to the transmit field inhomogeneity for TESS quantification of T1, the B1+ sequence will be implemented together during the scan. As for the in vivo experiments, 5 healthy subjects were recruited.
\ue3\ue3In our phantom experiments we observe time-varying B_0 drifts, which causes T1 and T2 quantification bias. In this study, a 3rd-order polynomial regression is utilized to correct the phase of the k-space raw data. Our Bland-Altman analysis shows a good agreement in T1 and T2 quantification on our phantom, using reference and TESS scans. Our in vivo results show a much reduced cerebrospinal fluid flow artifacts when radial sampling is employed. We are able to image a whole brain in about 7.5 minutes and calculate the whole brain T1 and T2 maps
Exploring the Clustering Phenomenon of Forex Extreme Illiquidity Events Using Mixture Models and Hidden Markov Models
No full textThis study conducts an in-depth investigation into the fluctuations of extreme liquidity values in the foreign exchange market. We improved the Amihud Liquidity Index and integrated it with mixture models and Hidden Markov Models for our analysis. Through the mixture models, we found that the market's extreme liquidity values follow a mixed distribution composed of multiple Poisson distributions, and that liquidity remains relatively stable for the most part. However, when we introduced the Hidden Markov Model, it revealed a more profound and accurate way of understanding market liquidity fluctuations, highlighting its strong advantages in predicting and depicting these changes. The Hidden Markov Model suggests that the frequency of market state changes is not high, indicating that market participants' behavior is relatively conservative. They make investment decisions based on their expectations for the future, rather than simply based on the current market state. However, during certain periods, such as when confronted with significant economic news or policy changes, market liquidity can undergo very drastic changes. In comparing models, we found that despite the Hidden Markov Model having fewer parameters, it is more potent in explaining market liquidity fluctuations, demonstrating its superiority in analyzing extreme liquidity values in the foreign exchange market. The findings of this study hold significant reference value for understanding changes in foreign exchange market liquidity and for investors in formulating their investment strategies
A Hessian Estimator on Planar Grids
No full textIn this thesis, we demonstrate a generalization of classical calculus theorems to functions defined on a graph. When the graph is a 2D grid, we propose a new definition of the {\it Hessian} {\it operator} and compare it with Moriguchi-Murota's {\it Hessian} {\it operator}. For verifying our {\it Hessian} {\it operator}, we construct a polynomial that is corresponded to . We show that, if , then our {\it Hessian} , and the sum of eigenvalues of is the well-known graph Laplacian of .
Moreover, we discuss the validity of the Bochner formula, which holds on Riemannian manifolds, on the 2D grid. We are able to show that the Bochner formula holds for homogeneous quadratic polynomials on grids with zero curvatur
Cryptocurrency Market Risk Assessment and Investment Strategy Research Based on Factor Analysis
No full textSince the 2020 pandemic, the Federal Reserve's quantitative easing and investors' low-cost funds spurred a surge in stocks and cryptocurrency. However, 2022's inflation rise and subsequent interest rate hikes led to decreased asset values, notably causing a loss of about $1.5 trillion in cryptocurrency market value.
This research aims to help investors understand the risks and features of the cryptocurrency market. We used Barra's EUE3 model to construct a multi-factor risk model considering factors such as size, momentum, volatility, liquidity, and network centrality. We analyzed the returns of eligible cryptocurrencies from 2017 to 2022, finding that the market value-weighted model accurately predicted returns by nearly 90%.
In practice, we implemented a core-satellite investment strategy and found that cryptocurrencies with smaller size, poorer liquidity, and median network centrality performed best. Through this approach, we aim to comprehend cryptocurrency market risks and establish effective investment strategies