學術產出-學位論文
文章檢視/開啟
書目匯出
-
題名 集成式模型應用於預測院內心臟驟停：整合多維度生命徵象數據
Bagged Ensemble Model for In-Hospital Cardiac Arrest Prediction: Integrating Multi-Dimensional Vital Signs Data作者 翁祐祥
Weng, Yu-Hsiang貢獻者 邱淑怡
Chiu, Shu-I
翁祐祥
Weng, Yu-Hsiang關鍵詞 心臟驟停
數據不平衡
集成式學習
相互資訊
cardiac arrest
imbalanced data
ensemble learning
mutual information日期 2024 上傳時間 4-九月-2024 15:02:34 (UTC+8) 摘要 在本研究中，我們探討了如何利用生命徵象 (vital signs) 的時序資料來預測住院病人的心臟驟停 (In-Hospital Cardiac Arrest)。我們將生命徵象的時序數據轉換為三維的特徵資料，以便更有效地捕捉時間序列中的變化模式。為了提升特徵選擇的準確性，我們採用了相互資訊 (mutual information) 作為選擇指標，以確保選出的特徵對於預測模型具有最重要的貢獻。 面對數據不平衡 (imbalanced data) 的挑戰，我們運用了各種資料擴增方法來平衡樣本分佈，從而提高模型的泛化能力和預測準確性。接著，我們應用了多種機器學習算法來建立預測模型，並利用集成式學習 (ensemble learning) 方法進行模型的整合。這些方法的結合旨在充分發揮不同模型的優勢，提高預測的穩定性和準確性。 實驗結果顯示，經過特徵選擇、資料擴增以及集成式學習等一系列處理後，我們所建立的模型在預測住院病人心臟驟停的準確性上均取得了顯著的提升。本研究的結果不僅展示了時序數據在預測心臟驟停方面的潛力，也為未來在此領域的研究提供了新的思路和方法。
In this study, we investigated how to utilize time-series data of vital signs to predict in-hospital cardiac arrest. We transformed the time-series data of vital signs into three-dimensional feature data to more effectively capture the variation patterns within the time series. To enhance the accuracy of feature selection, we employed mutual information as a selection criterion to ensure that the chosen features made the most significant contribution to the predictive model. In addressing the challenge of imbalanced data, we used various data augmentation methods to balance the sample distribution, thereby improving the model's generalization capability and predictive accuracy. Subsequently, we applied multiple machine learning algorithms to build the prediction model and employed ensemble learning methods for model integration. The combination of these methods aimed to fully leverage the strengths of different models, enhancing prediction stability and accuracy. Experimental results demonstrated that after feature selection, data augmentation, and ensemble learning, the model we developed achieved significant improvements in the accuracy of predicting in-hospital cardiac arrest. The findings of this study not only highlight the potential of time-series data in predicting cardiac arrest but also provide new insights and methods for future research in this area.參考文獻 1. Smith, G.B., et al., A Comparison of the Ability of the Physiologic Components of Medical Emergency Team Criteria and the U.K. National Early Warning Score to Discriminate Patients at Risk of a Range of Adverse Clinical Outcomes*. Critical Care Medicine, 2016. 44(12). 2. Subbe, C.P., et al., Validation of a modified Early Warning Score in medical admissions. QJM: An International Journal of Medicine, 2001. 94(10): p. 521-526. 3. Chang, H.K., et al. Early Detecting In-Hospital Cardiac Arrest Based on Machine Learning on Imbalanced Data. in 2019 IEEE International Conference on Healthcare Informatics (ICHI). 2019. 4. Tanii, R., et al., Impact of dynamic parameter of trends in vital signs on the prediction of serious events in hospitalized patients -a retrospective observational study. Resuscitation Plus, 2024. 18: p. 100628. 5. Kwon, J.m., et al., An Algorithm Based on Deep Learning for Predicting In‐Hospital Cardiac Arrest. Journal of the American Heart Association, 2018. 7(13): p. e008678. 6. Su, C.-F., et al., Improved inpatient deterioration detection in general wards by using time-series vital signs. Scientific Reports, 2022. 12(1): p. 11901. 7. Mahajan, P., et al. Ensemble Learning for Disease Prediction: A Review. Healthcare, 2023. 11, DOI: 10.3390/healthcare11121808. 8. Pei, C., et al., Ensemble Learning for Early-Response Prediction of Antidepressant Treatment in Major Depressive Disorder. Journal of Magnetic Resonance Imaging, 2020. 52(1): p. 161-171. 9. Awad, A., et al., Early hospital mortality prediction of intensive care unit patients using an ensemble learning approach. International Journal of Medical Informatics, 2017. 108: p. 185-195. 10. Su, C.F., et al., Improved inpatient deterioration detection in general wards by using time-series vital signs. Sci Rep, 2022. 12(1): p. 11901. 11. Vergara, J.R. and P.A. Estévez, A review of feature selection methods based on mutual information. Neural Computing and Applications, 2014. 24(1): p. 175-186. 12. Ross, B.C., Mutual information between discrete and continuous data sets. PloS one, 2014. 9(2): p. e87357. 13. Mohammed, R., J. Rawashdeh, and M. Abdullah. Machine Learning with Oversampling and Undersampling Techniques: Overview Study and Experimental Results. in 2020 11th International Conference on Information and Communication Systems (ICICS). 2020. 14. Chawla, N.V., et al., SMOTE: synthetic minority over-sampling technique. Journal of artificial intelligence research, 2002. 16: p. 321-357. 15. He, H., et al. ADASYN: Adaptive synthetic sampling approach for imbalanced learning. in 2008 IEEE international joint conference on neural networks (IEEE world congress on computational intelligence). 2008. Ieee. 16. Kumari, S., D. Kumar, and M. Mittal, An ensemble approach for classification and prediction of diabetes mellitus using soft voting classifier. International Journal of Cognitive Computing in Engineering, 2021. 2: p. 40-46. 17. Salur, M.U. and İ. Aydın, A soft voting ensemble learning-based approach for multimodal sentiment analysis. Neural Computing and Applications, 2022. 34(21): p. 18391-18406. 描述 碩士
國立政治大學
資訊科學系
111753226資料來源 http://thesis.lib.nccu.edu.tw/record/#G0111753226 資料類型 thesis dc.contributor.advisor 邱淑怡 zh_TW dc.contributor.advisor Chiu, Shu-I en_US dc.contributor.author (作者) 翁祐祥 zh_TW dc.contributor.author (作者) Weng, Yu-Hsiang en_US dc.creator (作者) 翁祐祥 zh_TW dc.creator (作者) Weng, Yu-Hsiang en_US dc.date (日期) 2024 en_US dc.date.accessioned 4-九月-2024 15:02:34 (UTC+8) - dc.date.available 4-九月-2024 15:02:34 (UTC+8) - dc.date.issued (上傳時間) 4-九月-2024 15:02:34 (UTC+8) - dc.identifier (其他 識別碼) G0111753226 en_US dc.identifier.uri (URI) https://nccur.lib.nccu.edu.tw/handle/140.119/153393 - dc.description (描述) 碩士 zh_TW dc.description (描述) 國立政治大學 zh_TW dc.description (描述) 資訊科學系 zh_TW dc.description (描述) 111753226 zh_TW dc.description.abstract (摘要) 在本研究中，我們探討了如何利用生命徵象 (vital signs) 的時序資料來預測住院病人的心臟驟停 (In-Hospital Cardiac Arrest)。我們將生命徵象的時序數據轉換為三維的特徵資料，以便更有效地捕捉時間序列中的變化模式。為了提升特徵選擇的準確性，我們採用了相互資訊 (mutual information) 作為選擇指標，以確保選出的特徵對於預測模型具有最重要的貢獻。 面對數據不平衡 (imbalanced data) 的挑戰，我們運用了各種資料擴增方法來平衡樣本分佈，從而提高模型的泛化能力和預測準確性。接著，我們應用了多種機器學習算法來建立預測模型，並利用集成式學習 (ensemble learning) 方法進行模型的整合。這些方法的結合旨在充分發揮不同模型的優勢，提高預測的穩定性和準確性。 實驗結果顯示，經過特徵選擇、資料擴增以及集成式學習等一系列處理後，我們所建立的模型在預測住院病人心臟驟停的準確性上均取得了顯著的提升。本研究的結果不僅展示了時序數據在預測心臟驟停方面的潛力，也為未來在此領域的研究提供了新的思路和方法。 zh_TW dc.description.abstract (摘要) In this study, we investigated how to utilize time-series data of vital signs to predict in-hospital cardiac arrest. We transformed the time-series data of vital signs into three-dimensional feature data to more effectively capture the variation patterns within the time series. To enhance the accuracy of feature selection, we employed mutual information as a selection criterion to ensure that the chosen features made the most significant contribution to the predictive model. In addressing the challenge of imbalanced data, we used various data augmentation methods to balance the sample distribution, thereby improving the model's generalization capability and predictive accuracy. Subsequently, we applied multiple machine learning algorithms to build the prediction model and employed ensemble learning methods for model integration. The combination of these methods aimed to fully leverage the strengths of different models, enhancing prediction stability and accuracy. Experimental results demonstrated that after feature selection, data augmentation, and ensemble learning, the model we developed achieved significant improvements in the accuracy of predicting in-hospital cardiac arrest. The findings of this study not only highlight the potential of time-series data in predicting cardiac arrest but also provide new insights and methods for future research in this area. en_US dc.description.tableofcontents 第一章 緒論 1 第二章 文獻探討 4 第一節 預測院內心臟驟停 4 第二節 集成式模型醫療數據的應用 6 第三章 研究方法 8 第一節 實驗數據 8 第二節 數據前處理 8 第三節 模型使用 10 第四節 模型結果整合 12 第五節 評估方法 13 第四章 實驗分析 16 第一節 資料集特徵擴增 16 第二節 訓練集與測試集 18 第三節 特徵選擇標準 19 第四節 特徵群選擇策略一：單一特徵 23 第五節 特徵群選擇策略二：變量方向篩選 25 第六節 比較整合模型與擴增方法 27 第七節 相關文獻方法的比較 31 第五章 結論 36 參考文獻 (References) 38 zh_TW dc.format.extent 1103558 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0111753226 en_US dc.subject (關鍵詞) 心臟驟停 zh_TW dc.subject (關鍵詞) 數據不平衡 zh_TW dc.subject (關鍵詞) 集成式學習 zh_TW dc.subject (關鍵詞) 相互資訊 zh_TW dc.subject (關鍵詞) cardiac arrest en_US dc.subject (關鍵詞) imbalanced data en_US dc.subject (關鍵詞) ensemble learning en_US dc.subject (關鍵詞) mutual information en_US dc.title (題名) 集成式模型應用於預測院內心臟驟停：整合多維度生命徵象數據 zh_TW dc.title (題名) Bagged Ensemble Model for In-Hospital Cardiac Arrest Prediction: Integrating Multi-Dimensional Vital Signs Data en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) 1. Smith, G.B., et al., A Comparison of the Ability of the Physiologic Components of Medical Emergency Team Criteria and the U.K. National Early Warning Score to Discriminate Patients at Risk of a Range of Adverse Clinical Outcomes*. Critical Care Medicine, 2016. 44(12). 2. Subbe, C.P., et al., Validation of a modified Early Warning Score in medical admissions. QJM: An International Journal of Medicine, 2001. 94(10): p. 521-526. 3. Chang, H.K., et al. Early Detecting In-Hospital Cardiac Arrest Based on Machine Learning on Imbalanced Data. in 2019 IEEE International Conference on Healthcare Informatics (ICHI). 2019. 4. Tanii, R., et al., Impact of dynamic parameter of trends in vital signs on the prediction of serious events in hospitalized patients -a retrospective observational study. Resuscitation Plus, 2024. 18: p. 100628. 5. Kwon, J.m., et al., An Algorithm Based on Deep Learning for Predicting In‐Hospital Cardiac Arrest. Journal of the American Heart Association, 2018. 7(13): p. e008678. 6. Su, C.-F., et al., Improved inpatient deterioration detection in general wards by using time-series vital signs. Scientific Reports, 2022. 12(1): p. 11901. 7. Mahajan, P., et al. Ensemble Learning for Disease Prediction: A Review. Healthcare, 2023. 11, DOI: 10.3390/healthcare11121808. 8. Pei, C., et al., Ensemble Learning for Early-Response Prediction of Antidepressant Treatment in Major Depressive Disorder. Journal of Magnetic Resonance Imaging, 2020. 52(1): p. 161-171. 9. Awad, A., et al., Early hospital mortality prediction of intensive care unit patients using an ensemble learning approach. International Journal of Medical Informatics, 2017. 108: p. 185-195. 10. Su, C.F., et al., Improved inpatient deterioration detection in general wards by using time-series vital signs. Sci Rep, 2022. 12(1): p. 11901. 11. Vergara, J.R. and P.A. Estévez, A review of feature selection methods based on mutual information. Neural Computing and Applications, 2014. 24(1): p. 175-186. 12. Ross, B.C., Mutual information between discrete and continuous data sets. PloS one, 2014. 9(2): p. e87357. 13. Mohammed, R., J. Rawashdeh, and M. Abdullah. Machine Learning with Oversampling and Undersampling Techniques: Overview Study and Experimental Results. in 2020 11th International Conference on Information and Communication Systems (ICICS). 2020. 14. Chawla, N.V., et al., SMOTE: synthetic minority over-sampling technique. Journal of artificial intelligence research, 2002. 16: p. 321-357. 15. He, H., et al. ADASYN: Adaptive synthetic sampling approach for imbalanced learning. in 2008 IEEE international joint conference on neural networks (IEEE world congress on computational intelligence). 2008. Ieee. 16. Kumari, S., D. Kumar, and M. Mittal, An ensemble approach for classification and prediction of diabetes mellitus using soft voting classifier. International Journal of Cognitive Computing in Engineering, 2021. 2: p. 40-46. 17. Salur, M.U. and İ. Aydın, A soft voting ensemble learning-based approach for multimodal sentiment analysis. Neural Computing and Applications, 2022. 34(21): p. 18391-18406. zh_TW
