運用深度學習方法預測阿茲海默症惡化與腦中風手術存活

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：25

、訪客IP：18.226.93.207

姓名

吳振豪(Jhen-Hao Wu) 查詢紙本館藏

畢業系所

資訊工程學系在職專班

論文名稱

運用深度學習方法預測阿茲海默症惡化與腦中風手術存活
(Deep learning in predicting outcomes of Alzheimer′s Disease aggravation and MCA infarction survival)

相關論文

★ 整合深度學習方法預測年齡以及衰老基因之研究

★ 運用深度學習方法預測癌症種類及存活死亡與治癒復發

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

至系統瀏覽論文 ( 永不開放)

摘要(中)

近年深度學習對醫療領域有深遠的影響，深度學習結合醫療影像應用技術，且運用機器學習能力快速及精準地判斷大量的醫療影像數據，協助醫生提高疾病診斷的正確率，更進而以相關的病理資訊來預測分析疾病發生風險與機率，全球學者仍持續研究如何運用深度學習的能力應用在各項智慧醫療發展上大腦疾病複雜又難以處理，腦部疾病中的一類為失智症，失智症中最常見的為阿茲海默症(Alzheimer′s disease ,AD)，佔失智病患人數約50%-70%，目前尚無實證有效的藥物治療方式，因此目前研究方向朝向延緩病程的惡化進行。而腦中風為腦部疾病中常見的神經中樞疾病，腦中風患者在入院前的病理狀態影響手術後的存活機率，本篇論文以深度學習模型可對不同類型醫療數據資料進行執行推斷能力，來預測阿茲海默症之惡化與腦中風手術後存活率。
論文中實作了機器學習的隨機森林樹，梯度提升樹，支援向量機與深度學習的深度神經網路，並利用預測阿茲海默症惡化結果來進行效能比較，接者區分資料集為病人病理資訊與空汙資訊，試圖找出阿茲海默症惡化危險因素。另以相同模式建立模組，針對腦中風患者入院前的腦中風評量表，以評估病患各項生理指數，來了解是否對於手術後存活率有影響。

摘要(英)

In recent years, deep learning has a far-reaching impact on the medical field, deep learning combined with medical imaging application technology, and the use of machine learning ability to quickly and accurately judge a large number of medical imaging data, to help doctors improve the correct rate of disease diagnosis, and then with relevant pathological information to predict and analyze the risk and probability of disease occurrence, global scholars continue to study how to use the ability of deep learning in the development of intelligent medical.
Brain diseases are complex and difficult to deal with, one of the brain diseases is dementia, the most common of which is Alzheimer′s disease, accounting for about 50%-70% of the number of dementia patients, there is no proven and effective drug treatment, so the current research direction towards delaying the deterioration of the course of the disease. Stroke is a common neurocreatic disease in brain diseases, and the pathological state of stroke patients before admission affects the chances of survival after surgery, this paper uses the deep learning model to make inferences about different types of medical data to predict the deterioration of Alzheimer′s disease and the survival rate after stroke surgery. In this paper, the random forest of machine learning, Gradient Boosting Trees, SVM and deep learning are used to compare the effectiveness by predicting the deterioration results of Alzheimer′s disease, and the recipient distinguishes the data set for the patient′s pathological information and the air pollution information, in an attempt to find out the risk factors for the deterioration of Alzheimer′s disease. In the same model, modules were created to evaluate the patient′s physiological indices for the pre-hospital stroke evaluation of stroke patients to see if they had an impact on post-operative survival.

關鍵字(中)

★ 阿茲海默症
★ 深度學習
★ 機器學習

關鍵字(英)

★ Alzheimer′s
★ Gradient Boosting
★ SVM

論文目次

中文摘要 i
Abstract ii
圖目錄 iv
表目錄 v
章節目次 vi
第一章緒論 1
1.1 研究背景、動機及目的 1
1.2 論文架構 3
第二章相關研究及文獻探討 4
2.1 深度學習(Deep Learning) 4
2.1.1 感知機 4
2.1.2 深度神經網路Deep Neural Network(DNN) 5
2.1.3 梯度下降法(Gradient descent) 7
2.1.4 交叉熵損失函式(Binary Cross Entropy) 8
2.2 分類器(Class Classifier) 9
2.2.1 隨機森林(Random forest) 9
2.1.2 梯度提升樹(Gradient Boosting Trees) 10
2.1.3 支援向量機(Support Vector Machine, SVM) 12
2.3 機器學習應用於阿茲海默症(Alzheimer′s Disease) 13
2.4 阿茲海默症(Alzheimer′s Disease) 危險因素 15
2.4.1心血管疾病高風險因子 15
2.4.2空氣污染風險因子 16
第三章整體實驗架構與方法 18
3.1 深度學習網路開發套件 18
3.1.1 TensorFlow 18
3.1.2 Google Colaboratory 18
3.2 資料前處理 19
3.2.1 正規化(Normalization) 19
3.2.1 標準化(Standardization) 19
3.3驗證方法 20
3.3.1 F-Measure 20
3.3.2 ROC Curve and AUC 21
3.3.3 K-Fold Cross Validation 22
第四章實驗結果 24
4.1 阿茲海默症資料集(Alzheimer′s Disease Dataset) 24
4.2 腦中風資料集(MCA infarction Dataset) 31
4.3 實驗流程 34
4.3.1 阿茲海默症實驗 34
4.3.2 腦中風實驗 41
4.4 阿茲海默症模型資料分析 46
第五章結論及未來研究方向 47
5.1 結論 47
5.2 未來研究方向 48
參考文獻 49

參考文獻

1 Lecun, Y., Bengio, Y. & Hinton, G. Deep learning. Nature 521, 436-444, doi:10.1038/nature14539 (2015).
2 Tan, Y., Wang, J. & Zurada, J. M. J. I. t. o. n. n. Nonlinear blind source separation using a radial basis function network. 12, 124-134 (2001).
3 Collins, F. S. & Varmus, H. A New Initiative on Precision Medicine. New England Journal of Medicine 372, 793-795, doi:10.1056/nejmp1500523 (2015).
4 Lynch, C. J. A. s. & Dementia. World Alzheimer Report 2019: Attitudes to dementia, a global survey: Public health: Engaging people in ADRD research. 16, e038255 (2020).
5 Castellani, R. J., Rolston, R. K. & Smith, M. A. Alzheimer Disease. Disease-a-Month 56, 484-546, doi:10.1016/j.disamonth.2010.06.001 (2010).
6 Krenker, A., Bešter, J., Kos, A. J. A. N. N. M. A. & InTech, B. A. Introduction to the artificial neural networks. 1-18 (2011).
7 Song, H. A. & Lee, S.-Y. in International conference on neural information processing. 466-473 (Springer).
8 Sze, V., Chen, Y.-H., Yang, T.-J. & Emer, J. S. Efficient Processing of Deep Neural Networks: A Tutorial and Survey. Proceedings of the IEEE 105, 2295-2329, doi:10.1109/jproc.2017.2761740 (2017).
9 Liu, W. et al. A survey of deep neural network architectures and their applications. 234, 11-26 (2017).
10 De Boer, P.-T., Kroese, D. P., Mannor, S. & Rubinstein, R. Y. A Tutorial on the Cross-Entropy Method. Annals of Operations Research 134, 19-67, doi:10.1007/s10479-005-5724-z (2005).
11 Buja, A., Stuetzle, W. & Shen, Y. J. W. d., November. Loss functions for binary class probability estimation and classification: Structure and applications. 3 (2005).
12 Biau, G. & Scornet, E. A random forest guided tour. TEST 25, 197-227, doi:10.1007/s11749-016-0481-7 (2016).
13 Fisher, C. K., Smith, A. M. & Walsh, J. R. Machine learning for comprehensive forecasting of Alzheimer’s Disease progression. Scientific Reports 9, doi:10.1038/s41598-019-49656-2 (2019).
14 Kivipelto, M. et al. Obesity and vascular risk factors at midlife and the risk of dementia and Alzheimer disease. 62, 1556-1560 (2005).
15 Whitmer, R. A., Sidney, S., Selby, J., Johnston, S. C. & Yaffe, K. J. N. Midlife cardiovascular risk factors and risk of dementia in late life. 64, 277-281 (2005).
16 Reijmer, Y. D. et al. Development of Vascular Risk Factors over 15 Years in Relation to Cognition: The H oorn Study. 60, 1426-1433 (2012).
17 Virta, J. J. et al. Midlife cardiovascular risk factors and late cognitive impairment. 28, 405-416 (2013).
18 Knopman, D. et al. Cardiovascular risk factors and cognitive decline in middle-aged adults. 56, 42-48 (2001).
19 Moulton, P. V., Yang, W. J. J. o. e. & health, p. Air pollution, oxidative stress, and Alzheimer′s disease. 2012 (2012).
20 Qiu, H. et al. Machine learning approaches to predict peak demand days of cardiovascular admissions considering environmental exposure. BMC Medical Informatics and Decision Making 20, doi:10.1186/s12911-020-1101-8 (2020).
21 Goldsborough, P. A Tour of TensorFlow. arXiv pre-print server, doi:None
arxiv:1610.01178 (2016).
22 Bisong, E. 59-64 (Apress, 2019).
23 Perkel, J. M. Why Jupyter is data scientists’ computational notebook of choice. Nature 563, 145-146, doi:10.1038/d41586-018-07196-1 (2018).
24 Yu, Wei, G.-Y. & Brooks, D. Benchmarking TPU, GPU, and CPU Platforms for Deep Learning. arXiv pre-print server, doi:None
arxiv:1907.10701 (2019).
25 Hripcsak, G. Agreement, the F-Measure, and Reliability in Information Retrieval. Journal of the American Medical Informatics Association 12, 296-298, doi:10.1197/jamia.m1733 (2005).
26 Anguita, D., Ghelardoni, L., Ghio, A., Oneto, L. & Ridella, S. in ESANN. 441-446.

指導教授

王家慶許藝瓊(Jia-Ching Wang Yi-Chiung Hsu)

審核日期

2022-1-22

推文