34254 research outputs found
Sort by
Synthesis of hydrophobic segment-incorporated poly[2-(tert-butylamino)ethyl methacrylate] and the application in antibacterial nanofibers
No full textIn today's life, it is not uncommon for people to fight against various pathogenic microorganisms. From small skin abrasions to large surgical wounds, the crisis of bacterial infection exists at any time. The current common treatment method is to use small molecules as antibiotics to achieve bactericidal effect by preventing or destroying specific physiological functions of microorganisms. However, the impact on specific mechanisms can easily cause mutations in the genetic factors of microorganisms which form drug-resistant microorganisms and seriously weaken the original curative effect of antibiotics that make the affected area worse and even killing patients.
Therefore, macromolecular antibacterial agents such as antibacterial polymers began to gradually replace common antibiotic drugs. These polymers usually contain cationic groups and hydrophobic groups at the same time. Cations are used to locate and attach to anionic groups on the cell wall of microorganisms, while hydrophobic groups destroy cell walls and cell membranes. Hydrophobicity not only affects microorganisms, but also affects Mammalian cell membranes. Therefore, it is crucial to control the overall polymer hydrophobicity or amphiphilic balance to improve the selectivity of the polymer for microorganisms and ordinary cells.
The monomer 2-(tert-butylamino)ethyl methacrylate (TA) and the monomer n-butyl methacrylate (n-BMA) were added into the polymerization reaction bottle for atom transfer radical polymerization (ATRP) in order to synthesis poly[2-(tert-butylamino)ethyl methacrylate] (PTA) and poly[2-(tert-butylamino)ethyl methacrylate]-co-poly[n-butyl methacrylate] (PTAB) with different proportions of lipophilic segments. After the synthesis, we have analyzed the optimal chemical structure by biocompatibility test and antibacterial test with Staphylococcus aureus (S. aureus) and Escherichia coli (E. coli) of copolymer PTAB. Subsequently, PTAB with the optimal chemical structure is mixed with polylactide (PLA) and subjected to electrospinning process. Adjusting the electrospinning parameters to study the similar fiber diameter and different polymer mixing ratios, the difference in the antibacterial ability of the nanofibers was detected and the optimal mixing ratio was selected. PLA-PTAB nanofibers with the optimal mixing ratio were loaded with anti-inflammatory drugs to form a wound dressing with antibacterial and anti-inflammatory properties, and in vitro drug release simulations and animal experiments were carried out
Chronic Kidney Disease Detection from Electrocardiogram Using Convolutional Neural Networks
No full textThe kidney is an essential organ responsible for metabolizing waste, balancing acid-base levels, and regulating fluid in the body. However, despite advances in technology and medical care, the number of people affected by chronic kidney disease (CKD) continues to rise steadily, making it the leading cause of healthcare expenditure in Taiwan. As a result, CKD has become a significant public health issue. What makes CKD particularly alarming is the lack of obvious symptoms in its early stages, often leading to delayed diagnosis and treatment. Chronic kidney disease is commonly accompanied by various cardiovascular complications such as hypertension, left ventricular hypertrophy, and heart failure. These complications can influence the manifestations of electrocardiograms (ECGs). Additionally, kidney dysfunction, due to its vital role, can lead to electrolyte imbalances, metabolic abnormalities, and factors that affect ECG readings, such as hyperkalemia and hypocalcemia. Therefore, ECG serves as a non-invasive method to assess cardiovascular disease in CKD patients, playing a crucial role in early detection and treatment of cardiovascular complications. In this study, we focused on ECG analysis and employed deep learning techniques to extract relevant features for predicting the presence of chronic kidney disease in patients. We proposed two deep learning models for CKD prediction. The first model utilized a two-dimensional convolutional neural network to extract features from 12-lead ECG signals and determine whether a patient has CKD. To further leverage the available data, the ECG signals were digitized and subjected to discrete wavelet transform to remove noise before further analysis. The experimental results demonstrated that the proposed models achieved accuracy rates of 85.33% and 84.65%, as well as area under the curve (AUC) values of 93.05% and 92.5% on the testing data, respectively. Overall, the deep learning models presented in this study effectively predict chronic kidney disease from ECG signals, demonstrating significant clinical utility and providing valuable insights into the application of deep learning in the field of medicine
The research of the literariness and geographic features of 'Xu\ue4\ub8Xiake' travel notes'
No full textThis thesis explores the literature characteristics and geographical attributes of the work XU Xiake\ue2s Travels. Xu Xiake was born in the years of the Wanli emperor\ue2s reign, due to the political instability, the rise in commerce and the popularity of tourism in his time he decided to abandon the original thought of participating in the imperial examinations \uef\ubcalso known as the civil service exams\uef\ubc, resolving to travelling throughout China. He documents his travels through writing a descriptive diary, although at first glance it may seem as a regular geographical guide book on travel, it retains its literature properties and demonstrates incredible literary techniques. Hence, Xu Xiake\ue2s travels remains equally balanced in both its geographical and literature characteristics. In terms of its literature elements, the Xu Xiake\ue2s Travels integrates the techniques of poet Liu Zongyuan, as seen in the great Tang dynasty poet\ue2s work \ue2Eight Travels of Yongzhou\ue2.Xu not only can make vivid descriptions of the spectacular views he witnessed, but also reflect his ideology and sentiments combined with the nature he witnessed, moreover he also possesses the expertise of interweaving long and short verses creating a rhythmic transition between his lines when read by his readers. On the other hand, Xu also conserves his accurate encounters of the geographical environment in his travels that includes meticulous details for example the height of the mountains, flowing direction of the rivers, distances and locations of his destinations, ecology, distribution of mineral resources and local traditions. In addition, this has led to it being inducted to the Siku Quanshu under the subcategory of geography in the larger category of history.
As of the main structure, if we want to conduct academic research on the geographical and historical properties of Xu Xiake\ue2s Travels. We will need to focus on
the historical background and his family to gain insight. Hence, we can trace the
societal backgrounds of the late Ming dynasty to discover that leisure travelling was common in his belonging epoch. His time frequently saw intellects composing their voyages into diaries like his in which this trend furtherly peaked in the late Ming dynasty. As from the historical records we can clearly see that Xu Xiake was born in a supportive affluent family along with his mother\ue2s aid providing him generous funds. Above all concluding that Xu Xiake authorship was contributed by society and family influence.
Xu Xiake\ue2s Travels utilizes its literal tone of voice to record geographical information and simultaneously the literature must be upheld by its geographical details. Instead of the traditional from top to bottom approach used by traditional Chinese authors, Xu Xiake\ue2s Travels implements a mutually inclusive form of writing in which the geography and literature benefits each other. Xu Xiake\ue2s Travel has the quantity and quality of content never foreseen by anyone, including its invaluable literary and geographical composition, earning it the title of \ue2the Zenith of traveling diaries\ue2.
In conclusion, this thesis approaches this topic from focusing on the Late Ming dynasty era, analyzing the connection between the geographical and literary attributes of Xu Xiake\ue2s Travel that also includes comparing it with other travel diaries in order to emphasize its specialty among its counterparts
Quantification of Blood Water T2 using a T2-prepared bSSFP Sequence: Measurement and Affecting Variables
No full textMagnetic Resonance Imaging (MRI) is a non-invasive imaging technique that utilizes changes in magnetic fields to visualize various tissues in the human body. Magnetic Resonance (MR) images provide detailed information about different anatomical structure, and with appropriate processing, can be used to estimate physiologic parameters, such as blood oxygen saturation level (HbO2). Among many different MR-based techniques for HbO2 measurements, a T2-based method (T2-Oxymetry, T2O) relies on a series of T2-weighted images for estimation of blood T2 relaxation time. MR images accurate estimation of T2 values is crucial for assessing blood oxygenation, as errors in T2 estimation can lead to inaccurate blood oxygenation measurements, thereby affecting clinical diagnosis. Therefore, the objective of this study is to enhance the precision of blood T2 measurements by reducing factors that may affect the accuracy of T2 value estimation.
Two primary factors influence T2 measurements: the Signal-to-Noise Ratio (SNR) and the partial volume effect. A high SNR ensures reliable signal fitting results, which in turn improves the accuracy of T2 measurements. The partial volume effect combines the signals of the observed blood with those of surrounding tissues. Reducing this effect can decrease its impact on T2 values.
This study employed a specialized custom MRI scanning sequence that included background suppression, T2 magnetization preparation, and Balanced Steady-State Free Precession (bSSFP). bSSFP is renowned for its high SNR and T2/T1 contrast. However, under the condition of field inhomogeneity, banding artifacts might appear. This research investigates the accuracy of bSSFP in capturing and estimating blood T2 values.
Phantom and human experiments were conducted to evaluate the effect of adjusting various parameters, including bSSFP encoding schemes, excitation angles, and background suppression configuration, on image results and T2 values. For bSSFP acquisition, excitation angles were adjust ed between 10 and 60 degrees. bSSFP encoding schemes included Centric Mode, where signal collection begins at the k-space center, and Linear Mode, where it begins at the k-space edge. Background suppression adjustments compared SSSS and SSNN configurations against images without background suppression.
Results revealed that adjusting the bSSFP excitation angle showed that images with a 60-degree angle had a high SNR, while those with a 10-degree angle had a lower SNR. Adjusting the excitation angle had minimal effect on T2 quantification results. However, images with a higher SNR showed greater precision in T2 quantification. The encoding schemes, in bSSFP affected T2 measurement results: The Linear Mode, having a longer T1 recovery time from the end of T2 magnetization preparation to the capture of the central k-space signal, resulted in brighter background signals. In contrast, the Centric Mode immediately collected central k-space signals after T2 magnetization preparation, reducing prolonged background signal recovery. Without background suppression, the tissue signals in the images were stronger, making the fitting process in the region of interest (ROI) more challenging.
In conclusion, this study examined the effects of bSSFP excitation angles, scanning sequence encoding schemes, and background suppression adjustments on T2 measurement values, aiming to improve the precision of blood T2 values. The experiments highlighted that the best precision in T2 measurement was achieved using a 60-degree excitation angle, Centric Mode encoding, and background suppression. Future research can incorporate other variables like blood flow rate and slice thickness to further enhance T2 measurement precision
A Preliminary Study to Analyze the Growth Experiences of Deviants:Children\ue2s Perspectives on Parental Education
No full textIn recent years, there has been a surge in fraud cases, which has led to a decline in social morality. As a researcher, I have a history of 10 years acting in anti-social ways and 15 years assuming the role of a pastoral counselor to continuously guide and support deviants after I reintegrated into society. Throughout the counseling process, I have observed that all the cases have been significantly influenced by their parental upbringing experiences during formative years in the family environment. The purpose of this research is to explore, from the perspective of the children, the parental upbringing experiences of deviants during their developmental journey. It aims to examine how a lack of proper understanding and skills in child-rearing among parents can increase the risk of children becoming deviants in their adulthood. The study employed a qualitative research approach using in-depth interviews and focus group interviews. A total of 13 adults were interviewed. The transcriptions of the in-depth interviews and the records of the focus group interviews were compiled, analyzed, and synthesized to derive the research findings. Research Findings: 1. Through the investigation of parental upbringing in the families of deviants, it was found that: (a) Parents generally care for their children, but there is room for improvement in their parenting approaches. (b) Parental upbringing is influenced by the previous generation. (c) Parental upbringing in the family environment affects attachment relationships. 2. The impact of negative parental upbringing in the families of deviants on their children's deviant behaviors. 3. The negative psychological effects on the children of deviants due to negative parental upbringing in their families. 4. The challenges faced by deviants upon reintegrating into society. The study recommends that due to the significant impact of parental upbringing in the family environment on the lifelong development of children, the government should strengthen the implementation of parental education policies under the existing Family Education Law. This enhancement aims to promote stable parental relationships within families, thereby reducing the risk and likelihood of individuals becoming deviant. The study also provides recommendations for preventive policies in parental education, policies regarding the reintegration of deviants into society, and suggestions for primary caregivers
Using Deep Learning to Predict Hyperkalemia from Single Lead ECG, a Transfer Learning Approach for Personalized Medicine
No full textIntroduction
Hyperkalemia, as one of the most common metabolic abnormalities seen in critical
patients will causes cardiotoxic effects and has been associated with a series of
Electrocardiogram (ECG) abnormalities that eventually lead to arrythmia and cardiac
arrest. In clinical practice, the sensitivity of physician readers in the ECG diagnosis of
hyperkalemia has been estimated to be as low as 34% to 43% due to its various
appearance. This condition often resulted in misdiagnosis, delayed medical treatment,
and increased mortality risk of critical ill patients.
Deep learning algorithm allows computer to learn directly from data and has achieved
great progress on assisting diagnosis of cardiac arrythmia, myocardial infarction on ECG.
Several studies had focused on developing deep learning model to recognize dyskalemia
from ECG but none of them demonstrated satisfactory prediction accuracy on single lead
ECG, which was used as continuous monitoring in intensive care unit (ICU). Managing
critical ill patients is a race against time. Early detection of ECG change on hyperkalemia
is crucial to enhance differential diagnosis process. With the assistance from AI, we
believe physicians are able to diagnose early, to provide treatment more efficiently, and
reduce unnecessary blood test on daily practice.
The purpose of this study is to develop and validate a deep learning model to recognize
hyperkalemia from ambulatory ECG monitor in ICU.
Method
Data for this study was generated from Medical Information Mart from Intensive Care
(MIMIC-III), an openly available database developed by Massachusetts Institute of
Technology. We included patients admitted to ICU with multiple serum potassium test
results, and matched ECG data from MIMIC-III database. Lead II ECG from ambulatory
monitor were collected and labeled as hyperkalemia and normal regarding their
corresponding potassium level. We matched and segmented ECG data based on the
timestamp of serum potassium test.
A one-dimensional convolution neural network based deep learning model is first
developed to predict hyperkalemia in a generic population from MIMIC-III database.
Once the model achieved a state-of-art performance, it was then utilized in an active
transfer learning process to perform patient adaptive heartbeat classification tasks.
Result
The results show that by acquiring a few data from each new patient, the personalized
model can improve the accuracy of hyperkalemia detection significantly from an average
of 0.604 \uc2\ub1 0.211 to 0.980 \uc2\ub1 0.078 when compared with the generic model. The Area
Under the receiver operating characteristic Curve level also improved from 0.729 \uc2\ub1 0.240
to 0.945 \uc2\ub1 0.094.
Conclusion
By utilizing deep transfer learning method, we were able to build a clinical standard
model for hyperkalemia detection from ambulatory ECG monitor. These findings could
potentially be extended to applications that continuously monitor one'
Adaptation of Digital Marketing in Timely Customer Behavior Changed due to Covid-19 Pandemic (Shopivch Case)
No full textThis research purpose is to investigate the change in Shopivch\ue2s customer behavior due to the impact of the COVID-19 pandemic in Indonesia. Also, investigating the digital marketing strategy and the brand extension strategy could be effective in facing the chaotic situation caused by the virus. The approach of this study uses a quantitative approach, with interviews with the owner of the company and 4 of the employees. Data is collected by investigating the company\ue2s internal and external factors that influence their business and then using SWOT analysis to find the best possible strategy for the company. The findings of the interviews indicate that the threat score is more than the opportunity score, and that the company's strength score is greater than its vulnerability score. According to the cartesian diagram, the best possible strategy is the strength-threats strategy that emphasizes the diversification strategy. At this position, indicates the company suffered a lot of threats from the external environment but had quite significant internal strength. Due to unpredictable economic conditions, the first strategy is to maintain the high-quality product with an affordable price. An attractive discount also can attract new prospective customers and sustain the loyalty of the customer. Last, use the brand extension strategy based on the customer's needs
First-principles calculation and machine learning prediction of three-dimensional covalent organic frameworks
No full textWe used the data from the first-principles calculations to train a machine learning model that can replace the first-principles calculations, overcoming the problem of large calculations in the first-principles calculations for large systems. After verifying that the accuracy of the model and the results of first-principles calculations have very little error, we used the model to test the stability of 52 three-dimensional covalent organic framework structures, including calculating stable structures, formation energies, condensation energies, \uce Point oscillation frequency and molecular dynamics. In addition, we also calculated the elastic coefficients of 52 three-dimensional covalent organic frameworks, and judged the mechanical stability of their structures. Using only the DFT calculation results of small molecules, we can predict the structure and energy of large systems with a much smaller amount of calculation, and the accuracy is close to the results of DFT. This paper demonstrates the possibility of using machine learning methods to study the structure and properties of materials with many atoms
A Study on the Factors Affecting the Return of Real Estate Investment Trusts \uef\ubcREITs\uef\ubcin the United States
No full textThis study focuses on publicly traded real estate investment trusts (REITs) in the United States and covers the period from January 2001 to April 2022. The study uses the least squares method to conduct regression analysis and explore the explanatory power of different factors on the returns of REITs. Among the Fama-French three-factor model, Carhart four-factor model, and Fama-French five-factor model, all types of REITs, except health REITs, which are considered a basic necessity industry, show a significant positive relationship between excess returns and market risk premium factors. Meanwhile, retail, office, residential, health, and industrial REITs, which are considered more traditional and conservative industries, show a significant positive relationship with the investment factor (CMA). In addition, OTC-listed stocks bring significant excess premiums, and the Covid-19 pandemic has led to a structural change in the use of real estate, resulting in an increase in excess returns for REITs.
Further analysis of company factors reveals that larger REITs may dilute their returns by choosing more investment targets, leading to lower stock returns. When REITs increase their financial leverage, they are considered more risky targets, leading to a decrease in stock prices. Profitability, on the other hand, contributes to stock performance. In addition, REIT investors prefer companies with tangible assets such as real estate and higher stock returns in the past year, while they are less likely to favor companies with high dividend payout ratios leading to less retained earnings for future investments, which may limit future cash flows. Lastly, the factors that affect the returns of REITs vary depending on the characteristics of each investment type
An Efficient TIPN-Table-Based Algorithm for Mining Top-k High On-shelf Utility Itemsets with Positive/Negative Profits Under the Consideration of the Local/Global Minimum Count
No full textHigh Utility Itemset Mining (HUIM) utilizes the threshold value to extract the High Utility Itemset (HUI) from the transactional database. However, it is hard to define a suitable threshold value, since it depends on the domain knowledge of the application. If the threshold value is set excessively high, only a limited number of HUIs are extracted. On the other hand, if the threshold value is set excessively low, a significant number of HUIs are extracted. Both of the above cases are not useful for making decision. Therefore, Top-k High Utility Itemset Mining (Top-k HUIM) solves the problem of setting a suitable threshold. The user can define a k value, which represents the number of HUIs. Moreover, there may exist some itemsets occurring at a specific time interval, which have a chance to become HUI. Since the tranditional HUIM algorithm does not consider the transaction with the time interval, we can not apply the HUIM algorithm directly. Thus, High On-shelf Utility Itemset Mining (HOUIM) addresses the above problem. The proportion of the utility value of the item in all of the time intervals, when the itemset appears is used for determining wether the itemset is a High On-shelf Utility Itemset (HOUI) or not. Furthermore, the tranditional HUIM does not consider the item with the negative profit, which happens in the real life. In the Top-k High On-shelf Utility Itemset Mining (Top-k HOUIM), the KOSHU algorithm utilizes the utility list based data structure and ignores the item with the negative profit in the process of overestimating the utility of the itemset. Thus, the KOSHU algorithm needs the less memory and processing time. However, the KOSHU algorithm has to scan the database two times and sort the database one time. Therefore, in this thesis, we propose an efficient algorithm based on the TIPN_Table to mines Top-k HOUIs from the static database. Our proposed data structures include TIPN_Table, MINC_Table, IO_Bitmap and TIUL. In the TIPN_Table, we record positive items, positive utilities, negative items and negative counts. MINC_Table is used for storing the local/global minimum counts of all of the items with negative profits. On the other hand, we propose the TWUGC value to improve the RTWU value to decrease the overestimated utility, which uses the global minimum count of all of the items with negative profits. In our algorithm, we only need to scan the database once to construct the TIPN_Table. From our performance study, we have shown that our proposed algorithm is more efficient than the KOSHU algorithm