結合 PSO 及 LSTM 於預測服務處理的時間— 使用急診室中風治療案例研究

簡易檢索 / 詳目顯示

回結果列表

研究生：	Vicki Ismi Caneca Vicki Ismi Caneca
論文名稱：	結合 PSO 及 LSTM 於預測服務處理的時間— 使用急診室中風治療案例研究 Integrating PSO and LSTM to Predict Service Processing Time – Using Emergency Room Stroke Treatment Case Study
指導教授：	歐陽超 Chao Ou-Yang
口試委員:	郭人介 Ren-Jieh Kuo 羅士哲 Shih-Che Lo
學位類別：	碩士 Master
系所名稱：	管理學院 - 工業管理系 Department of Industrial Management
論文出版年：	2021
畢業學年度：	109
語文別：	英文
論文頁數：	67
中文關鍵詞：	健康照護研究、健康預測、全域粒子群演算法、長短期記憶模型
外文關鍵詞：	Health Service Research, Health Forecasting, GPSO, LSTM
相關次數：	點閱：168 下載：5
分享至:	分享至facebook 分享至twitter

查詢本校圖書館目錄查詢臺灣博碩士論文知識加值系統勘誤回報

上一筆

健康照護研究(Health Service Research, HSR) 是一新興領域的產業，且在我們的生活中扮演著重要的角色，包含了醫院的臨床服務、藥物、醫療設備、與醫療保健相關的支持服務，也在全國、全球的GDP中佔有一席之地。在健康照護研究領域中已有多種方法及技術被提出討論，如預測演算法、各類技術使用、軟體開發、風險分析、緩解措施等改善服務的方法。此外，也有研究開發了「健康預測」以提升醫療保健服務行業的服務及其功能。健康預測改善了醫院的治療、預測存活率、更具經濟效益及具有其他優勢。
本研究使用健康預測的概念，並聚焦於預測台灣當地醫院急診室中各項醫療活動的處理時間以進行中風疾病治療。預測處理時間可以用來將醫院提供的治療表現進行分類並將效益最佳化。
本研究藉由進階最佳化演算法得到資料集分類的最佳化，並且結合另一預測演算法作為變數。首先，使用傳統的集群分析(KMeans Clustering)用來獲得活動群集數量，並得到七個群集。再使用全域粒子群演算法(Global Particle Swarm Optimization, GPSO)在每個群集的中心點最佳化得到每個粒子之間最短的距離。最後使用處理時間及每個活動群集的編號建立機器學習LSTM演算法預測模型。本研究中使用方均根誤差(Root Mean Square Error, RMSE)評估預測模型的準確率，並得到26%的預測錯誤。在未來的研究中需要將額外的參數輸入訓練，以及獲得更多可用的資料，都能幫助改善預測模型準確率。

Health Service Research (HSR) is an emerging field industry that plays an essential part in our lives. It includes the clinical services from a hospital, drugs, medical equipment, healthcare-related support services, and a role in national and global GDP. Several methodologies and approaches have been used and discussed along with several purposes in HSR, such as prediction algorithm, technology-used, software development, risk analysis, mitigations, and others to improve the services. In addition, researchers have developed so-called health forecasting to improve the Health care Service Industry for both services and their features. Health forecasting improves such treatment in a hospital, predicting survivability rate, economic-wise, and other advantages.
In this research, health forecasting is used and focused on predicting each activity's processing time in a Taiwan local hospital emergency room for a stroke disease treatment. Predicting the processing time can be used to optimize the utilization tools and categorize the performance of the treatment provided by the hospital.
Categorization of the dataset is used and optimized by an advanced optimization algorithm to be added as another variable for the prediction algorithm used in this research. First, Traditional KMeans Clustering is used to obtain the number of cluster activities, which resulted in seven clusters. The position of each centroid obtained from the KMeans clustering is optimized by selected Global Particle Swarm Optimization (GPSO) to get minimum distance with each particle. Finally, the processing time and the number of each activity cluster are used to build a prediction model using LSTM machine learning algorithm. The prediction is evaluated using Root Mean Square Error (RMSE), which resulted in 26 minutes of prediction error. Additional variables as the input and data availability are needed to improve the prediction performance and benefit for future research.

ABSTRACT    i
摘要    ii
ACKNOWLEDGMENT    iii
CONTENTS    iv
LIST OF FIGURES    vi
LIST OF TABLES    vii
CHAPTER 1 INTRODUCTION    1
1    Background    1
2     Purpose    3
3    Research Scope and Limitation    3
4     Organization of the Thesis    3
CHAPTER 2 LITERATURE REVIEW    4
1    Stroke    4
2    Health Forecasting    4
3    Data Categorization    7
4    KMeans Clustering    7
5    RNN-LSTM    7
6    Optimization Algorithm    9
6.1    Particle Swarm Optimization (PSO)    10
6.2    Global Particle Swarm Optimization (GPSO)    10
7    Literature Review    12
CHAPTER 3 RESEARCH METHODOLOGY    17
1    Data Collection    18
2    Data Preprocessing    18
2.1    Data Categorization    19
3    Prediction Model    19
4    Model Evaluation    20
5    Optimization    20
CHAPTER 4 IMPLEMENTATION    22
implementation    22
1    Data Collection    22
2    Dataset Overview    22
2.1    Data Preprocessing    25
2.1    Data Categorization    30
3    Processing Time Prediction    41
3.1    Parameter Settings    41
3.2    LSTM Prediction to Predict Processing Time    42
3.3    Evaluating LSTM Model    43
4    Data Categorization Analysis    50
5    Prediction Model Analysis    51
6    Condition for LSTM Model    52
CHAPTER 5 CONCLUSION AND FUTURE RESEARCH    53
1    Summary    53
2    Future Research    54
REFERENCES    55

                                

[1] Economywatch, “Health Care Industry,” p. 1, 2010.
[2] K. N. Lohr and D. M. Steinwachs, “Health Service Research: An Evolving Definition of the Field,” Health Services Research: Impacting Health Practice and Policy Through State-of-the-Art Research and Thinking, pp. 15-17, 2002.
[3] N. Arakawa and C. Anderson, “Challenges and opportunities in conducting health services research through international collaborations: A review of personal experiences,” Research in Social and Administrative Pharmacy, pp. 1609-1613, 2020.
[4] I. N. Soyiri and D. D. Reidpath, “An overview of health forecasting,” Environment Health and Preventive Medicine, pp. 1-9, 2012.
[5] H. Xiao, X. Jiang, C. Chen, W. Wang, C.-Y. Wang, A. A. Ali, A. Berthe, R. K. Moussa and V. Diaby, “Using time series analysis to forecast the health-related quality of life of post-menopausal women with non-metastatic ER+ breast cancer: A tutorial and case study,” Research in Social and Administrative Pharmacy, pp. 1095-1099, 2020.
[6] J.-w. Lee, H.-s. Lim, D.-w. Kim, S.-a. Shin, J. Kim, B. Yoo and K.-h. Cho, “The development and implementation of stroke risk prediction model in National Health Insurance Service's personal health record,” Computer Methods and Programs in Biomedicine, pp. 253-257, 2018.
[7] M. Francesconi, A. Minichino, R. Carrion, R. D. Chiaie, A. Bevilacqua, M. Parisi, S. Rullo, F. S. Bersani, M. Biondi and K. Cadenhead, “Psychosis prediction in secondary mental health services. A broad, comprehensive approach to the "at risk mental state" syndrome,” European Psychiatry, pp. 96-104, 2017.
[8] National Institute of Neurological Disorders and Stroke, “Stroke: Hope Through Research,” Patient & Caregiver Education, p. 1, 2020.
[9] M. A. Bartek, R. C. Saxena, S. Solomon, C. T. Fong, L. D. Behara, R. Venigandla, K. Velagapudi, B. G. Nair and J. D. Lang, “Improving Operation Room Efficiency: A machine learning approach to predict case-time duration,” American College of Surgeons, pp. 346-354, 2019.
[10] S. Barnes, E. Hamrock, M. Toerper, S. Siddiqui and S. Levin, “Real-time prediction of inpatient length of stay for discharge prioritization,” US National Library of Medicine National Institutes of Health, pp. 2-10, 2016.
[11] World Health Organization, “The top 10 causes of death,” WHO, 2020.
[12] US National Library of Medicine National Institutes of Health, “An overview of health forecasting,” Environmental Health and Preventive Medicine, pp. 1-9, 2013.
[13] D. F. Hernandez-Suarez, S. Ranka, Y. Kim, A. Latib, J. Wiley, A. Lopez-Candales, D. S. Pinto, M. C. Gonzalez, H. Ramakrishna, C. Sanina, B. G. Nieves-Rodriguez, J. Rodriguez-Maldonado, R. F. Maldonado, I. J. Rodriguez-Ruiz, I. d. L. Sant'Ana, K. A. Wiley, P. Cox-Alomar, P. A. Villablanca and A. Roche-Lima, “Machine-Learning-Based In-Hospital Mortality Prediction for Transcatheter Mitral Valve Repair in the United States,” Cardiovascular Revascularization Medicine, pp. 22-28, 2021.
[14] J. H. Park, S. D. Shin, K. J. Song, K. J. Hong, K. J. Ro, Y. S. Ro, J.-W. Choi and S. W. Choi, “Prediction of good neurological recovery after out-of-hospital cardiac arrest: A machine learning analysis,” Resuscitation, pp. 127-135, 2019.
[15] Y. Gautam, “Transfer Learning for COVID-19 cases and deaths forecast using LSTM network,” ISA Transactions, pp. 39-51, 2021.
[16] S. Jiang, K.-S. Chin, G. Qu dan K. L. Tsui, “An integrated machine learning framework for hospital readmission prediction,” Knowledge-Based Systems, pp. 73-90, 2018.
[17] I. Dabbura, “K-means Clustering: Algorithm, Applications, Evaluation Methods, and Drawbacks,” 18 September 2018. [Online]. Available: https://towardsdatascience.com/k-means-clustering-algorithm-applications-evaluation-methods-and-drawbacks-aa03e644b48a. [Accessed 22 May 2021].
[18] J. T. Connor, R. D. Martin and I. Member, “Recurrent Neural Networks and Robust Time Series Prediction,” IEEE Transactions on Neural Networks, vol. 5, pp. 240-254, 1994.
[19] A. Tharwat dan W. Schenck, “A conceptual and practical comparison of PSO-style optimization algorithms,” Expert Systems with Applications, vol. 167, 2021.
[20] J. Kennedy, R. C. Eberhart and Y. Shi, “Chapter seven - The Particle Swarm,” Swarm Intelligence, pp. 287-325, 2001.
[21] S. A. Aoudia, E.-H. Guerrout and R. Mahiou, Medical Image Segmentation using Particle Swarm Optimization, Paris: ESI, 2014.
[22] J. J. Jamian, M. N. Abdullah, H. Mokhlis, M. W. Mustafa dan A. H. A. Bakar, “Global Particle Swarm Optimization for High Dimension Numerical Functions Analysis,” Applied Mathematics, 2014.
[23] J. Wang, J. Cao and S. Yuan, “Shear wave velocity prediction based on adaptive particle swarm optimization optimized recurrent neural network,” Journal of Petroleum Science and Engineering, vol. 194, 2020.
[24] T. Gupta, “towardsdatascience,” 2019. [Online]. Available: https://towardsdatascience.com/data-preprocessing-in-data-mining-machine-learning-79a9662e2eb. [Accessed 31 May 2021].
[25] Y. Chauhan, “towardsdatascience,” 2020. [Online]. Available: https://towardsdatascience.com/k-means-clustering-with-python-code-explained-5a792bd19548. [Accessed 01 June 2021].
[26] S. Rengasamy and P. Murugesan, “PSO based data clustering with a different perception,” Swarm and Evolutionary Computation, vol. 64, Elsevier, 2021.
[27] M. A. Bartek, R. C. Saxena, S. Solomon, C. T. Fong, L. D. Behara, R. Venigandla, K. Velagapudi, B. G. Nair and J. D. Lang, “Improving Operating Room Efficiency: A machine learning approach to predict case-time duration,” US National Library of Medicine National Institutes of Health, pp. 346-354, 2019.
[28] X. Wei, L. Zhang, H.-Q. Yang, L. Zhang dan Y.-P. Yao, “Machine learning for pore-water pressure time-series prediction: Application of recurrent neural networks,” Geoscience Frontiers, vol. 12, no. 1, pp. 453-467, 2021.
[29] S. Rengasamy and P. Murugesan, “PSO based data clustering with a different perception,” Swarm and Evolutionary Computation, vol. 64, 2021.

簡易檢索 / 詳目顯示

相關論文