The application of machine learning algorithms for predicting length of stay before and during the COVID-19 pandemic: evidence from Wuhan-area hospitals DOI Creative Commons
Yang Liu,

Ruonan Liang,

Chengzhi Zhang

et al.

Frontiers in Digital Health, Journal Year: 2024, Volume and Issue: 6

Published: Dec. 13, 2024

Objective The COVID-19 pandemic has placed unprecedented strain on healthcare systems, mainly due to the highly variable and challenging predict patient length of stay (LOS). This study aims identify primary factors impacting LOS for patients before during pandemic. Methods collected electronic medical record data from Zhongnan Hospital Wuhan University. We employed six machine learning algorithms probability LOS. Results After implementing selection, we identified 35 variables affecting establish model. top three predictive were out-of-pocket amount, insurance, admission deplanement. experiments conducted showed that XGBoost (XGB) achieved best performance. MAE, RMSE, MAPE errors are lower than 3% average household registration in non-household Wuhan. Conclusions Research finds is reasonable predicting offers valuable guidance hospital administrators planning resource allocation strategies can effectively meet demand. Consequently, these insights contribute improved quality care wiser utilization scarce resources.

Language: Английский

Leveraging ChatGPT to optimize depression intervention through explainable deep learning DOI Creative Commons
Yang Liu,

Xingchen Ding,

Shun Peng

et al.

Frontiers in Psychiatry, Journal Year: 2024, Volume and Issue: 15

Published: June 6, 2024

Introduction Mental health issues bring a heavy burden to individuals and societies around the world. Recently, large language model ChatGPT has demonstrated potential in depression intervention. The primary objective of this study was ascertain viability as tool for aiding counselors their interactions with patients while concurrently evaluating its comparability human-generated content (HGC). Methods We propose novel framework that integrates state-of-the-art AI technologies, including ChatGPT, BERT, SHAP, enhance accuracy effectiveness mental interventions. generates responses user inquiries, which are then classified using BERT ensure reliability content. SHAP is subsequently employed provide insights into underlying semantic constructs AI-generated recommendations, enhancing interpretability Results Remarkably, our proposed methodology consistently achieved an impressive rate 93.76%. discerned always employs polite considerate tone responses. It refrains from intricate or unconventional vocabulary maintains impersonal demeanor. These findings underscore significance AIGC invaluable complementary component conventional intervention strategies. Discussion This illuminates considerable promise offered by utilization models realm healthcare. represents pivotal step toward advancing development sophisticated healthcare systems capable augmenting patient care counseling practices.

Language: Английский

Citations

7
Injecting new insights: How do review sentiment and rating inconsistency shape the helpfulness of airline reviews? DOI
Yang Liu,

Lihua Ma,

Yue Dou

et al.

Information Processing & Management, Journal Year: 2025, Volume and Issue: 62(4), P. 104088 - 104088

Published: Feb. 7, 2025

Language: Английский

Citations

0
Understanding COVID-19 vaccine hesitancy of different regions in the post-epidemic era: A causality deep learning approach DOI Creative Commons
Yang Liu, Chenxu Zhao, Chengzhi Zhang

et al.

Digital Health, Journal Year: 2024, Volume and Issue: 10

Published: Jan. 1, 2024

Objective This paper aims to understand vaccine hesitancy in the post-epidemic era by analyzing texts related reviews and public attitudes toward three prominent brands: Sinovac, AstraZeneca, Pfizer, exploring relationship of with prevalence epidemics different regions. Methods We collected 165629 Twitter user comments associated brands. The were labeled based on willingness attitude vaccination. utilize a causality deep learning model, Bert multi-channel convolutional neural network (BertMCNN), predict users’ mutually. Results When applied provided dataset, proposed BertMCNN model demonstrated superior performance traditional machine algorithms other models. It is worth noting that after March 2022, was more hesitant about Sinovac vaccines. Conclusions study reveals connection between epidemic analytical results obtained from this method can assist governmental health departments making informed decisions regarding vaccination strategies.

Language: Английский

Citations

1
How broadband internet shapes household tourism decisions: a double/debiased machine learning-based difference-in-difference approach DOI
Haowen Jia

Journal of Hospitality and Tourism Technology, Journal Year: 2024, Volume and Issue: unknown

Published: Oct. 18, 2024

Purpose This study aims to examine the influence of broadband internet on Chinese households’ tourism decisions and spending patterns. It reveals transformative potential digital connectivity in enhancing activity. Design/methodology/approach Using a rigorous double/debiased machine learning-based difference-in-difference (DMLDiD) method extensive panel data, this research quantitatively analyzes impact household engagement financial allocations. incorporates comprehensive robustness checks, including placebo tests algorithm variations, ensure validity findings. Findings Within households with access, results indicate significant increase 3.54% 31.24% participation tourism-related expenditures, respectively. attributes these outcomes enhanced incomes, facilitated online transactions alleviated credit constraints, highlighting notable disparities across urban versus rural settings among distinct demographic categories. Additionally, moderating effects marital status size reveal that married greater number members tend leverage access more effectively for making expenditures. Originality/value By pioneering application DMLDiD approach behaviors toward tourism, contributes novel insights economic discourse role infrastructure development. offers empirical evidence strategic implications policymakers industry professionals who seek enhance tourism.

Language: Английский

Citations

1
Clustering Analysis of Hotel Network Reviews Based on Text Mining Method DOI
Yao Wang, Fuguo Liu, Guodong Li

et al.

Industry science and engineering., Journal Year: 2024, Volume and Issue: 1(4), P. 51 - 59

Published: April 1, 2024

With the development of information technology, users use online platforms to post real-time comments express their preferences and opinions on goods or services. Online review expresses users' behavioral habits special preferences. In depth, analysis hotel reviews can improve adaptability services user needs. Effective mining vast data will provide value for tourism industry. Using text methods process data, multiple clustering were compared analyzed positive negative feature words from perspective experience. It was found that k-means++ algorithm had a better effect network achieved segmentation evaluation information. Unsupervised be used further classify comment into categories based reviews, providing intellectual support improving precision personalized service quality hotels.

Language: Английский

Citations

0
Stock movement prediction in a hotel with multimodality and spatio-temporal features during the Covid-19 pandemic DOI Creative Commons
Yang Liu, Lili Ma

Heliyon, Journal Year: 2024, Volume and Issue: 10(21), P. e40024 - e40024

Published: Nov. 1, 2024

The COVID-19 pandemic has underscored the importance of accurate stock prediction in tourism industry, particularly for hotels. Despite growing interest leveraging consumer reviews performance forecasting, existing methods often need to integrate rich, multimodal data from these fully. This study addresses this gap by developing a novel deep learning model, Multimodal Spatio-Temporal Graph Convolutional Neural Network (MSGCN), specifically designed predict hotel performance. Unlike traditional models, MSGCN captures spatial relationships between hotels using graph convolutional network and integrates information—including text, images, ratings reviews—into process. Our research builds on literature validating efficacy improving introducing spatio-temporal component that enhances accuracy. Through rigorous testing two diverse datasets, our model demonstrates superior compared approaches, showing robustness during after pandemic. findings provide valuable insights managers consumers, offering powerful tool making informed business decisions rapidly evolving market.

Language: Английский

Citations

0
The application of machine learning algorithms for predicting length of stay before and during the COVID-19 pandemic: evidence from Wuhan-area hospitals DOI Creative Commons
Yang Liu,

Ruonan Liang,

Chengzhi Zhang

et al.

Frontiers in Digital Health, Journal Year: 2024, Volume and Issue: 6

Published: Dec. 13, 2024

Objective The COVID-19 pandemic has placed unprecedented strain on healthcare systems, mainly due to the highly variable and challenging predict patient length of stay (LOS). This study aims identify primary factors impacting LOS for patients before during pandemic. Methods collected electronic medical record data from Zhongnan Hospital Wuhan University. We employed six machine learning algorithms probability LOS. Results After implementing selection, we identified 35 variables affecting establish model. top three predictive were out-of-pocket amount, insurance, admission deplanement. experiments conducted showed that XGBoost (XGB) achieved best performance. MAE, RMSE, MAPE errors are lower than 3% average household registration in non-household Wuhan. Conclusions Research finds is reasonable predicting offers valuable guidance hospital administrators planning resource allocation strategies can effectively meet demand. Consequently, these insights contribute improved quality care wiser utilization scarce resources.

Language: Английский

Citations

0