Search

Your search keyword '"Pan, Ke"' showing total 2,079 results
Start Over Author "Pan, Ke"
2,079 results on '"Pan, Ke"'

2. Applying Machine Learning Models with An Ensemble Approach for Accurate Real-Time Influenza Forecasting in Taiwan: Development and Validation Study

Author

Cheng, Hao-Yuan, Wu, Yu-Chun, Lin, Min-Hau, Liu, Yu-Lun, Tsai, Yue-Yang, Wu, Jo-Hua, Pan, Ke-Han, Ke, Chih-Jung, Chen, Chiu-Mei, Liu, Ding-Ping, Lin, I-Feng, and Chuang, Jen-Hsiang

Subjects
Computer applications to medicine. Medical informatics, R858-859.7, Public aspects of medicine, RA1-1270
Abstract

BackgroundChangeful seasonal influenza activity in subtropical areas such as Taiwan causes problems in epidemic preparedness. The Taiwan Centers for Disease Control has maintained real-time national influenza surveillance systems since 2004. Except for timely monitoring, epidemic forecasting using the national influenza surveillance data can provide pivotal information for public health response. ObjectiveWe aimed to develop predictive models using machine learning to provide real-time influenza-like illness forecasts. MethodsUsing surveillance data of influenza-like illness visits from emergency departments (from the Real-Time Outbreak and Disease Surveillance System), outpatient departments (from the National Health Insurance database), and the records of patients with severe influenza with complications (from the National Notifiable Disease Surveillance System), we developed 4 machine learning models (autoregressive integrated moving average, random forest, support vector regression, and extreme gradient boosting) to produce weekly influenza-like illness predictions for a given week and 3 subsequent weeks. We established a framework of the machine learning models and used an ensemble approach called stacking to integrate these predictions. We trained the models using historical data from 2008-2014. We evaluated their predictive ability during 2015-2017 for each of the 4-week time periods using Pearson correlation, mean absolute percentage error (MAPE), and hit rate of trend prediction. A dashboard website was built to visualize the forecasts, and the results of real-world implementation of this forecasting framework in 2018 were evaluated using the same metrics. ResultsAll models could accurately predict the timing and magnitudes of the seasonal peaks in the then-current week (nowcast) (ρ=0.802-0.965; MAPE: 5.2%-9.2%; hit rate: 0.577-0.756), 1-week (ρ=0.803-0.918; MAPE: 8.3%-11.8%; hit rate: 0.643-0.747), 2-week (ρ=0.783-0.867; MAPE: 10.1%-15.3%; hit rate: 0.669-0.734), and 3-week forecasts (ρ=0.676-0.801; MAPE: 12.0%-18.9%; hit rate: 0.643-0.786), especially the ensemble model. In real-world implementation in 2018, the forecasting performance was still accurate in nowcasts (ρ=0.875-0.969; MAPE: 5.3%-8.0%; hit rate: 0.582-0.782) and remained satisfactory in 3-week forecasts (ρ=0.721-0.908; MAPE: 7.6%-13.5%; hit rate: 0.596-0.904). ConclusionsThis machine learning and ensemble approach can make accurate, real-time influenza-like illness forecasts for a 4-week period, and thus, facilitate decision making.

Published
2020
Full Text
View/download PDF

23. Empirical validation of existing human carbon monoxide (CO) exposure models with suggested improvements relevant to groups in the general population

Author

Pan, Ke-Ting

Abstract

Carbon monoxide (CO) poisoning is an important public health issue globally. Several CO exposure models have been built to predict the kinetics of CO uptake and elimination in CO poisoning but based on a limited dataset confined to a handful of young, healthy male volunteers. It is important to improve these models by using a wider range of individuals to seek an optimal performance across a realistic range of characteristics of the general population. Therefore, to expand the validity and practical applicability of CO exposure models to people with different characteristics, such as age, sex, height, weight, smoking status and ethnicity, we carried out three studies to test the relationship between these factors and pulmonary function, represented by the Diffusion capacity of the Lung for CO (DLCO), as well as the rate of CO uptake and CO elimination. Specifically, we first explored how demographic, physiological and behavioural factors affect CO uptake and CO elimination among young volunteers. The experiment indicated that smoking status did not influence CO uptake and elimination in this group. Then, we collected pulmonary function test (PFT) data from two hospitals, one in the UK and one in Taiwan. Given the exceptional circumstances of the COVID-19 pandemic in the UK, obtaining ethics approval for hospital data was challenging, however with great effort, PFT data were collected from both hospitals. We found that males had a higher value of DLCO than females, and DLCO was positively associated with height and weight, and negatively associated with age. Later, the estimated DLCO results from PFT in relatively healthy individuals were used to update established Coburn-Forster-Kane (CFK) models. These showed that, for example, when the CO exposure scenario was the same, CO uptake and elimination rate were higher in a younger male than an older male. In conclusion, the updated model has the advantage of being able to predict CO uptake and elimination for a wider range of individuals compared to previous CFK models, factoring in their age, sex, height, weight, ethnic group, and smoking status. This could help estimate past CO exposure, and help medical staff recognize CO exposure and design optimal treatments for CO exposure victims.

Published
2022

24. Total synthesis of (+)-trichostatins A and C via the O-directed dialkylacetylene free radical hydrostannation reaction

Author

Pan, Ke and Stevenson, Paul

Subjects
total synthesis, Trichostatin C, O-directed free radical hydrostannation
Abstract

(+)-Trichostatin A (TSA) is a naturally-occurring histone deacetylase (HDAC) inhibitor of type I, II, and IV HDACs that exhibits powerful anticancer effects. (+)-Trichostatin C (TSC) is natural glucopyranosyl hydroxylamine derivatives of TSA with O-β-D-glucopyranosylhydroxylamine. We became interested in developing an enantioselective total synthesis of the (+)-TSA and (+)-TSC molecule to potentially show off the significant utility of the O-directed free radical hydrostannation of disubstituted alkynes for constructing trisubstituted alkenes in target molecules that contain a dialkylaniline ring system. Such a synthetic approach to (+)-TSA might potentially be adaptable to the creation of novel analogues that could be screened for their HDAC inhibitory effects, which are of potential therapeutic relevance to the treatment of diseases ranging from cancer right through to SARS-CoV-2 (COVID-19). Our new total syntheses of (+)-trichostatins A and C have been completed in high optical purity (97% ee), so confirming that the O-directed free radical hydrostannation with Ph3SnH/cat. Et3B can be successfully applied in dialkylaniline-containing dialkyl acetylene systems. With this new route, many novel avenues of the trichostatins can potentially be prepared and screened.

Published
2022

30. Multi-Party Dual Learning

Author

Gong, Maoguo, Gao, Yuan, Xie, Yu, Qin, A. K., Pan, Ke, and Ong, Yew-Soon

Subjects
Computer Science - Machine Learning
Abstract

The performance of machine learning algorithms heavily relies on the availability of a large amount of training data. However, in reality, data usually reside in distributed parties such as different institutions and may not be directly gathered and integrated due to various data policy constraints. As a result, some parties may suffer from insufficient data available for training machine learning models. In this paper, we propose a multi-party dual learning (MPDL) framework to alleviate the problem of limited data with poor quality in an isolated party. Since the knowledge sharing processes for multiple parties always emerge in dual forms, we show that dual learning is naturally suitable to handle the challenge of missing data, and explicitly exploits the probabilistic correlation and structural relationship between dual tasks to regularize the training process. We introduce a feature-oriented differential privacy with mathematical proof, in order to avoid possible privacy leakage of raw features in the dual inference process. The approach requires minimal modifications to the existing multi-party learning structure, and each party can build flexible and powerful models separately, whose accuracy is no less than non-distributed self-learning approaches. The MPDL framework achieves significant improvement compared with state-of-the-art multi-party learning methods, as we demonstrated through simulations on real-world datasets., Comment: submitted to IEEE Transactions on Cybernetics

Published
2021

40. Association between adherence to the Dietary Approaches to Stop Hypertension diet and serum uric acid

Author

Jie Feng, Yuchai Huang, Haozhuo Wang, Chao Wang, Hongbin Xu, Pan Ke, Yan He, Qingfeng Tian, Shiyi Cao, and Zuxun Lu

Subjects
Medicine, Science
Abstract

Abstract To explore the relationship between Dietary Approaches to Stop Hypertension (DASH) diet and serum uric acid (SUA) levels among the Chinese adult population, and verify the mediating effect of BMI between DASH diet and SUA levels. A total of 1125 adults were investigated using a self-administered food frequency questionnaire. SUA levels were determined by uricase colorimetry. The total DASH score ranged from 9 to 72. The relationship between the DASH diet and SUA levels was examined by multiple adjusted regression analysis. Method of Bootstrap was used to test the mediation effect of BMI in the correlation of the DASH diet and SUA levels. After multivariable adjustment, there was a significant linear relationship between the DASH diet and SUA (P

Published
2023
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results