臺灣博碩士論文加值系統

〈p align=justify〉〈font face=Times New Roman〉　　近年來，隨著電子化病歷的發展與普及，電子病歷資料量快速的增加。如何從電子病歷中萃取有價值的知識運用於協助醫療決策成為重要議題。為此，本研究提出了一系列新穎的疾病早期風險評估方法，透過資料探勘與深度學習技術分析大規模電子病歷的多種疾病風險因子，達成早期檢測慢性病風險之效益，並以多種慢性病為實例演繹。

　　首先，我們提出一項疾病風險關聯樣式探勘框架（Disease Risk Association Pattern Mining Framework，簡稱DR-APM）探勘電子病歷中隱含的疾病風險樣式，以檢驗病人是否具罹患特定慢性疾病之風險。DR-APM的技術要項包括疾病風險樣式探勘、基於風險樣式之早期風險評估、PubMed文獻檢索疾病風險樣式比對策略（Risk Patterns Matching in PubMed，簡稱RPM-PubMed）與統計分析。經RPM-PubMed實驗分析，DR-APM發現之風險模式可歸納為習知風險樣式與具新穎性之風險樣式。其中，疾病組與對照組的病人其主要風險樣式分佈的疾病類型具顯著差異，因此DR-APM應用於早期評估類風濕性關節炎風險可達成良好之準確率。

　　接著，為了提升疾病風險評估的效率與準確性，我們對電子病歷中存在大量疾病編碼屬性與稀疏矩陣的問題特性提出了一種融合了矩陣解構與機器學習的方法eDRAM (Early Disease Risk Assessment with the Matrix Factorization)。此方法以非負值矩陣分解算法，顯著的降低資料維度，重建新穎的風險因子以實行早期疾病風險評估方法。實驗結果顯示相較於所比較之疾病評估方法，eDRAM達成降低大量資料屬性、提升靈敏度 與評估速度。

　　深度學習網路中存在訓練模型運算耗時與耗費資源的問題。對此我們基於疾病編碼的週期時序性、概括型編碼與數量建構新穎風險特徵、並隨機抽樣降低訓練資料量與運用深度殘餘卷積神經網路，提出具延展性的時序概括型電子病歷深度學習（scalable Deep learning of Temporal generalized EHRs, 簡稱sDT-EHRs）。為評估有效性與泛用性，sDT-EHRs被應用於三種不同的慢性疾病並與三種先進的方法進行比較。實驗結果顯示sDT-EHRs表現出良好的可延展性，可基於相對少量病患資料所建構的模型，預測較大規模的病患資料且維持良好且相近的正確率。且sDT-EHRs在三種不同慢性疾病評估之準確率皆優於所比較之慢性疾病風險評估方法。

　　本研究基於現今精準醫療之需求，以資料探勘、機器學習與深度學習之方法為基石，系統化的探討與發展一系列疾病早期風險探勘與評估方法。為了運用真實世界大規模的醫病資料提升多種不同慢性疾病評估效益，我們設計了一系列實驗以評估所提出方法於效益與效率之進步性。本研究主要貢獻是探勘與建構新穎的風險因子模式及提升疾病早期評估方法的效益，其可提供醫學上進一步的驗證分析，且適合應用於評估不同的疾病，以此提升醫療效率。〈/font〉〈/p〉

〈p align=justify〉〈font face=Times New Roman〉　In recent years, the amount of electronic medical records (EMRs) has increased rapidly. Hence, obtaining valuable knowledge from EMRs to support medical decision making has become an important issue. To address this issue, in the thesis, we propose a set of novel early risk assessment methods for different chronic diseases by identifying diverse disease risk factors from the National Health Insurance Research Database (NHIRD).

　First, we propose a Disease Risk Association Pattern Mining Framework (DR-APM) to detect early risk for chronic diseases and rheumatoid arthritis was used as a case study. The main strategies of DR-APM include mining of disease risk pattern, associative classification, analysis with Risk Pattern Matching in PubMed (RPM-PubMed) and statistical analysis. The RPM-PubMed experiments show that the risk patterns discovered through DR-APM can be organized into well-known risk pattern type and potential novel risk pattern type. The experiments in statistical analysis reveal that there are significant differences in the disease categories of risk pattern distributions between the disease group and the control group. Based on the significant differences, DR-APM can achieve excellent accuracy in early risk assessment.

　Second, in order to deal with the problem of a large number of disease coding attributes and the sparse matrix problem in EMR database, we propose an early Disease Risk Assessment with the Matrix factorization method (eDRAM) that fuses machine learning and matrix factorization to identify latent risk factors from the EMR database. eDRAM uses a non-negative matrix decomposition algorithm to significantly reduce the data dimension and reconstruct novel risk factors for early disease risk assessment. The experiments demonstrate that eDRAM can reduce a large number of attributes and maintain better efficiency, stability and effectiveness compared to other state-of-the-art methods.

　Finally, in recent years, deep learning can achieve excellent performance in features recognition. However, the computational time-consuming and resource-intensive problems exist in the training model phase, especially dealing with large-scale attributes and data. To solve these problems to assess different types of diseases and improve accuracy, we propose an effective method called scalable Deep learning of Temporal generalized EHRs (sDT-EHRs). sDT-EHRs includes a novel temporal EHR representation model with an extraction algorithm, a random sampling method, and a deep residual convolutional neural network. To evaluate the effectiveness of sDT-EHRs for early risk assessment of multiple diseases, the following three chronic diseases: chronic obstructive pulmonary disease, systemic lupus erythematosus, and type 2 diabetes mellitus were assessed in the experiments, and sDT-EHRs was compared with state-of-the-art methods for early risk assessment of three chronic diseases via a large-scale nationwide medical database. Experimental evaluations of performance, scalability and applied to multiple chronic diseases yielded major three findings. First, this proposed EHR representation model is a combination of generalized disease codes that increase efficiency during the training phase. Second, sDT-EHRs outperforms other state-of-the-art methods during the risk assessment of the three chronic diseases. Finally, sDT-EHRs demonstrates good scalability to assess the diseases risk based on the disease models constructed from relatively small amounts of patient data and to maintain high performance when evaluating a large number of patients.

　This research mainly considers the needs of modern precision medical treatment, and systematically investigates and develops a set of early disease risk assessment frameworks based on the data mining, machine learning and deep learning techniques. In order to use real-world large-scale medical data for the early risk assessments of different chronic diseases, we design a set of experiments to evaluate the improvement of the proposed method in terms of efficiency and effectiveness. The main contribution of this study is to discover a variety of novel risk factors and improve the early risk assessment methods, which can provide further medical validation analysis and assessment of different diseases to improve medical care.〈/font〉〈/p〉

〈p align="justify"〉〈font face="Times New Roman"〉Contents

Abstract in Chinese I
Abstract III
Acknowledgement V
Contents VI
List of Tables IX
List of Figures X
Chapter 1 Introduction 1
1.1 Motivation 1
1.2 Overview of the Dissertation 3
1.2.1 Mining Risk Patterns from Nationwide Clinical Databases 5
1.2.2 Effective Early Risk Assessment with Matrix Factorization 6
1.2.3 Deep Learning for Early Risk Assessment 7
1.3 Organization of the Dissertation 8
Chapter 2 Background and Related Work 9
2.1 Data Mining in Early Assessment of Disease Risk 9
2.2 Data Mining in the Taiwan NHIRD 10
2.3 Early Assessment of Rheumatoid Arthritis 11
2.4 Non-negative Matrix Factorization for Disease Prediction 11
2.5 Deep Learning and Machine Learning Approaches for Disease Prediction 12
2.6 Summary 13
Chapter 3 Mining Disease Risk Patterns from Nationwide Clinical Databases for the Early Risk Assessment of Chronic Disease 14
3.1 Introduction 14
3.2 Methods 16
3.2.1 Framework and Workflow 16
3.2.2 Large-Scale Diagnostic Dataset 18
3.2.3 Statistical Analysis 19
3.2.4 Data Preprocessing 20
3.2.5 Risk Pattern Mining 21
3.2.6 10-Fold Cross-Validation 22
3.2.7 Well-known-degree Analysis of Risk Patterns 23
3.3 Results and Discussion 27
3.4.1 Risk Pattern Evaluations under Different Parameter Settings 27
3.4.2 Mining Significant Disease Risk Patterns 29
3.4.3 Distribution of Disease Risk Patterns in the ICD-9-CM Coding System 30
3.4.4 Significant RA Risk Patterns of Associative Classification 34
3.4 Summary 40
Chapter 4 Early Disease Risk Assessment with Matrix Factorization 42
4.1 Introduction 42
4.2 Methods 45
4.2.1 Overview of the Proposed Framework 45
4.2.2 Matrix Transformation 46
4.2.3 Discovery of Latent Risk Factors 49
4.2.4 Construction of Disease Risk Assessment Model 50
4.2.5 Disease Risk Assessment 50
4.2.6 Parameters 51
4.3 Materials 51
4.4 Experiments 53
4.4.1 Experimental Dataset 54
4.4.2 Experimental Environment 55
4.4.3 Experimental Measures 55
4.4.4 Experimental Results 56
4.4.5 Experimental Settings for Parameter R 56
4.4.6 Effectiveness Evaluation 58
4.4.7 Efficiency Evaluation 59
4.5 Discussion 60
4.6 Summary 61
Chapter 5 Deep Learning for Early Assessment of Chronic Diseases by Means of a Medical Database with Temporal Information 63
5.1 Introduction 63
5.2 Methods 65
5.2.1 Dataset 65
5.2.2 The Framework of the Proposed Method 67
5.2.3 Data Selection 68
5.2.4 TQDR Matrix Extraction 68
5.2.5 Temporal Deep Residual Learning 74
5.3 Experiments 76
5.3.1 The Experimental Dataset 76
5.3.2 Experimental Measures 77
5.3.3 Performance Evaluation 78
5.3.4 Scalability Evaluation 79
5.4 Summary 80
Chapter 6 Conclusions and Future Work 82
Bibliography 86

List of Tables

Table 3.1 Clinical course variables among patients categorized by RA diagnosis and gender. 20
Table 3.2 Characteristics of mined patterns. 30
Table 3.3 Risk pattern distribution between the RA group and non-RA group for all diseases categorized in ICD-9-CM. 31
Table 3.4 Sensitivity and specificity of the RA disease risk model with ten-fold cross-validation. 34
Table 3.5 RA risk patterns of autoimmune-related diseases. 37
Table 4.1 Baseline characteristics of RA patients in the cohort (2000–2008). 53
Table 4.2 Description of experimental data. 54
Table 4.3 Efficiency comparison on Dataset 3. 60
Table 5.1 Example of EHR with ICD-9 codes 64
Table 5.2 Baseline characteristics of patients in the cohort 66
Table 5.3 ICD-9 codes of the diseases included in the study 66
Table 5.4 First-level of ICD-9 codes hierarchy 70
Table 5.5 Sample size of dataset 1 for effectiveness evaluation 76
Table 5.6 Sample size for scalability evaluation 77
Table 5.7 Differences in the effectiveness based on sDT-EHRs across different data scales 80

List of Figures

Figure 1.1 Framework of the dissertation. 4
Figure 3.1 Workflow of data mining process for RA disease early detection. 17
Figure 3.2 Age distribution of cohort. 19
Figure 3.3 Timeline for data collection of RA group. 21
Figure 3.4 Risk Pattern Viewer. 25
Figure 3.5 Trend of RA assessment effects under different support thresholds. 27
Figure 3.6 Distribution of RA patients under different numbers of diagnosis records. 28
Figure 3.7 Trend of RA assessment effects under different diagnosis record numbers. 29
Figure 3.8 Trend in literature for single disease risk patterns in PubMed. 33
Figure 3.9 Trend in literature for PubMed pattern related mining. 33
Figure 4.1 Timeline for data collection and definition of RA patients. 44
Figure 4.2 Framework of proposed approach. 46
Figure 4.3 Flow chart of study cohort enrollment. 47
Figure 4.4 Example and concept of transformed patient–disease diagnosis matrix. 48
Figure 4.5 Using NMF to decompose the patient–disease diagnosis matrix. 50
Figure 4.6 10-fold cross–validation model 55
Figure 4.7 Effectiveness of RA risk assessment under different number of risk factors. 58
Figure 4.8 Comparison of the performance of eDRAM, CBS, CMAR, and BayesFM approaches. 59
Figure 5.1 Framework of the proposed disease risk assessment system 68
Figure 5.2 Observing and constructing diagnostic data using the time interval model 71
Figure 5.3 The algorithm of the proposed extraction of TQDR matrix 73
Figure 5.4 Example of a temporal quantitative disease risk matrix transformed from the EHRs 74
Figure 5.5 Architecture of the temporal deep residual learning network 75
Figure 5.6 Performance comparison of sDT-EHRs against D-EHRs, SVM, random forest, LDA, and LR for early assessment of COPD, T2DM, and SLE risks 79
Figure 5.7 Performance comparison of the scalability evaluation dataset and performance evaluation dataset for the early assessment of COPD and T2DM risks 79〈/font〉〈/p〉

(p align=justify)(font face=Times New Roman)[1]National Health Insurance Administration, Ministry of Health and Welfare, Executive Yuan. National health insurance annual report. Available: http://www1.nhi.gov.tw/Nhi_E-LibraryPubWeb/CustomPage/Periodical.aspx?FType=8
[2]R. Agrawal and R. Srikant, Fast Algorithms for Mining Association Rules in Large Databases, in Proceedings of the 20th International Conference on Very Large Data Bases, 1994, vol. 1215, pp. 487-499: Morgan Kaufmann Publishers Inc.
[3]M. R. Arbuckle, M. T. McClain, M. V. Rubertone, R. H. Scofield, G. J. Dennis, J. A. James et al., Development of Autoantibodies before the Clinical Onset of Systemic Lupus Erythematosus, New England Journal of Medicine, vol. 349, no. 16, pp. 1526-1533, 2003.
[4]Z. Bedran, C. Quiroz, J. Rosa, L. J. Catoggio, and E. R. Soriano, Validation of a Prediction Rule for the Diagnosis of Rheumatoid Arthritis in Patients with Recent Onset Undifferentiated Arthritis, International Journal of Rheumatology, vol. 2013, p. 548502, 2013, Art. no. 548502.
[5]M. Bergström, I. Ahlstrand, I. Thyberg, T. Falkmer, B. Börsbo, and M. Björk, ‘Like the worst toothache you’ve had’ – How people with rheumatoid arthritis describe and manage pain, Scandinavian Journal of Occupational Therapy, vol. 24, no. 6, pp. 468-476, 2017.
[6]M. W. Berry, M. Browne, A. N. Langville, V. P. Pauca, and R. J. Plemmons, Algorithms and applications for approximate nonnegative matrix factorization, Computational Statistics ＆ Data Analysis, vol. 52, no. 1, pp. 155-173, 2007.
[7]M. Z. Cader, A. Filer, J. Hazlehurst, P. de Pablo, C. D. Buckley, and K. Raza, Performance of the 2010 ACR/EULAR criteria for rheumatoid arthritis: comparison with 1987 ACR criteria in a very early synovitis cohort, Annals of the Rheumatic Diseases, vol. 70, no. 6, pp. 949-955, 2011.
[8]D. Cai, X. He, J. Han, and T. S. Huang, Graph Regularized Nonnegative Matrix Factorization for Data Representation, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 33, no. 8, pp. 1548-1560, 2011.
[9]B. Cao, D. Shen, J.-T. Sun, X. Wang, Q. Yang, and Z. Chen, Detect and track latent factors with online nonnegative matrix factorization, presented at the Proceedings of the 20th international joint conference on Artifical intelligence, Hyderabad, India, 2007.
[10]R. J. Carroll, A. E. Eyler, and J. C. Denny, Naïve Electronic Health Record Phenotype Identification for Rheumatoid Arthritis, AMIA Annual Symposium Proceedings, vol. 2011, pp. 189-196, 2011.
[11]R. J. Carroll, W. K. Thompson, A. E. Eyler, A. M. Mandelin, T. Cai, R. M. Zink et al., Portability of an algorithm to identify rheumatoid arthritis in electronic health records, Journal of the American Medical Informatics Association : JAMIA, vol. 19, no. e1, pp. 162-169, 2012.
[12]C.-C. Chang and C.-J. Lin, LIBSVM: A library for support vector machines, ACM Trans. Intell. Syst. Technol., vol. 2, no. 3, pp. 1-27, 2011.
[13]T.-F. Chao, C.-J. Liu, T.-C. Tuan, S.-J. Chen, T.-J. Chen, G. Y. H. Lip et al., Risk and Prediction of Sudden Cardiac Death and Ventricular Arrhythmias for Patients with Atrial Fibrillation – A Nationwide Cohort Study, Scientific Reports, vol. 7, p. 46445, 2017.
[14]R. Chatterjee, K. M. V. Narayan, J. Lipscomb, and L. S. Phillips, Screening adults for pre-diabetes and diabetes may be cost-saving, Diabetes care, vol. 33, no. 7, pp. 1484-1490, 2010.
[15]H. Y. Chen, Y. H. Lin, P. F. Thien, S. C. Chang, Y. C. Chen, S. S. Lo et al., Identifying Core Herbal Treatments for Children with Asthma: Implication from a Chinese Herbal Medicine Database in Taiwan, Evidence-Based Complementary and Alternative Medicine, vol. 2013, p. 125943, 2013, Art. no. 125943.
[16]P.-J. Chen, M.-C. Lin, M.-J. Lai, J.-C. Lin, H. H.-S. Lu, and V. S. Tseng, Accurate Classification of Diminutive Colorectal Polyps Using Computer-Aided Analysis, Gastroenterology, vol. 154, no. 3, pp. 568-575, 2018.
[17]Y.-F. Chen, C.-S. Lin, K.-A. Wang, L. O. A. Rahman, D.-J. Lee, W.-S. Chung et al., Design of a Clinical Decision Support System for Fracture Prediction Using Imbalanced Dataset, Journal of Healthcare Engineering, vol. 2018, p. 13, 2018, Art. no. 9621640.
[18]Y. Chen, Y. Li, R. Narayan, A. Subramanian, and X. Xie, Gene expression inference with deep learning, Bioinformatics (Oxford, England), vol. 32, no. 12, pp. 1832-1839, 2016.
[19]Y. Cheng, Y. Lin, K. Chiang, and V. S. Tseng, Mining Sequential Risk Patterns From Large-Scale Clinical Databases for Early Assessment of Chronic Diseases: A Case Study on Chronic Obstructive Pulmonary Disease, IEEE Journal of Biomedical and Health Informatics, vol. 21, no. 2, pp. 303-311, 2017.
[20]E. Choi, M. T. Bahadori, A. Schuetz, W. F. Stewart, and J. Sun, Doctor AI: Predicting Clinical Events via Recurrent Neural Networks, in Proceedings of the 1st Machine Learning for Healthcare Conference, Proceedings of Machine Learning Research, 2016, vol. 56, pp. 301-318: PMLR.
[21]E. Choy, Understanding the dynamics: pathways involved in the pathogenesis of rheumatoid arthritis, Rheumatology (Oxford), vol. 51, no. Suppl 5, pp. v3-v11, 2012.
[22]V. A. Cruz, L. Yamaguchi, C. N. Ribeiro, V. d. O. Magalhães, J. Rego, and N. A. d. Silva, Ulcerative colitis and rheumatoid arthritis: a rare association - case report, Revista Brasileira de Reumatologia, vol. 52, no. 4, pp. 648-650, 2012.
[23]D. K. D., O. D. C. I., H. Wolfgang, M. D. S., L. A. A., D. L. A. et al., The number of elevated cytokines and chemokines in preclinical seropositive rheumatoid arthritis predicts time to diagnosis in an age‐dependent manner, Arthritis ＆ Rheumatism, vol. 62, no. 11, pp. 3161-3172, 2010.
[24]X. Ding, Y. Zhang, T. Liu, and J. Duan, Deep learning for event-driven stock prediction, presented at the Proceedings of the 24th International Conference on Artificial Intelligence, Buenos Aires, Argentina, 2015.
[25]M. Dougados, M. Soubrier, A. Antunez, P. Balint, A. Balsa, M. H. Buch et al., Prevalence of comorbidities in rheumatoid arthritis and evaluation of their monitoring: results of an international, cross-sectional study (COMORA), Annals of the Rheumatic Diseases, vol. 73, no. 1, pp. 62-68, 2014.
[26]A. Ebringer and T. Rashid, Rheumatoid arthritis is caused by a Proteus urinary tract infection, APMIS, vol. 122, no. 5, pp. 363-368, 2014.
[27]N. B. Erichson, A. Mendible, S. Wihlborn, and J. N. Kutz, Randomized nonnegative matrix factorization, Pattern Recognition Letters, vol. 104, pp. 1-7, 2018.
[28]T. W. Fischer, H. I. Bauer, T. Graefe, U. Barta, and P. Elsner, Erythema multiforme-like drug eruption with oral involvement after intake of leflunomide, Dermatology, vol. 207, no. 4, pp. 386-389, 2004.
[29]P. Foti Daniela, M. Greco, E. Palella, and E. Gulletta, New laboratory markers for the management of rheumatoid arthritis patients, Clinical Chemistry and Laboratory Medicine (CCLM), vol. 52, no. 12, p. 1729, 2014.
[30]K. Fraser and L. Robertson, Chronic urticaria and autoimmunity, Skin therapy letter, vol. 18, no. 7, pp. 5-9, 2013.
[31]F. A. Gers, J. Schmidhuber, and F. Cummins, Learning to forget: continual prediction with LSTM, presented at the Ninth International Conference on Artificial Neural Networks (ICANN), 1999.
[32]A. Gibofsky, Overview of epidemiology, pathophysiology, and diagnosis of rheumatoid arthritis, The American Journal of Managed Care, vol. 18(13 Suppl), pp. S295-302, 2012.
[33]Y. P. M. Goekoop-Ruiterman, J. K. De Vries-Bouwstra, C. F. Allaart, D. Van Zeben, P. J. S. M. Kerstens, J. M. W. Hazes et al., Clinical and radiographic outcomes of four different treatment strategies in patients with early rheumatoid arthritis (the BeSt study): A randomized, controlled trial, Arthritis ＆ Rheumatism, vol. 52, no. 11, pp. 3381-3390, 2005.
[34]P. Groves, B. Kayyali, D. Knott, and S. Van Kuiken, The ‘big data’revolution in healthcare, McKinsey Quarterly, vol. 2, no. 3, 2013.
[35]B. Haikola, S. Huumonen, K. Sipilä, K. Oikarinen, T. Remes-Lyly, and A.-L. Söderholm, Radiological signs indicating infection of dental origin in elderly Finns, Acta Odontologica Scandinavica, vol. 71, no. 3-4, pp. 498-507, 2013.
[36]K. He, X. Zhang, S. Ren, and J. Sun, Deep Residual Learning for Image Recognition, in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 770-778.
[37]J. C. Ho, J. Ghosh, and J. Sun, Marble: high-throughput phenotyping from electronic health records via sparse nonnegative tensor factorization, presented at the Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining, New York, New York, USA, 2014.
[38]G. Hripcsak and D. J. Albers, Next-generation phenotyping of electronic health records, Journal of the American Medical Informatics Association : JAMIA, vol. 20, no. 1, pp. 117-121, 2013.
[39]T. W. J. Huizinga and A. H. M. van der Helmvan Mil, Prediction and prevention of rheumatoid arthritis, Revista Colombiana de Reumatología, vol. 14, pp. 106-114, 2007.
[40]J. Jaimes-Hernández, A. Mendoza-Fuentes, C. I. Meléndez-Mercado, and P. Aranda-Pereira, Chronic eosinophilic pneumonia: Autoimmune phenomenon or immunoallergic disease? Case report and literature review, Reumatología Clínica, vol. 8, no. 3, pp. 145-148, 2012.
[41]H. Jansen, C. Willenborg, W. Lieb, L. Zeng, P. G. Ferrario, C. Loley et al., Rheumatoid Arthritis and Coronary Artery Disease: Genetic Analyses Do Not Support a Causal Relation, The Journal of Rheumatology, vol. 44, no. 1, pp. 4-10, 2017.
[42]B. Jin, C. Che, Z. Liu, S. Zhang, X. Yin, and X. Wei, Predicting the Risk of Heart Failure With EHR Sequential Data Modeling, IEEE Access, vol. 6, pp. 9256-9261, 2018.
[43]S. Kapoor, Beyond rheumatoid arthritis: The close association between interstitial cystitis and Sjogren's syndrome, Neurourology and Urodynamics, vol. 34, no. 1, p. 101, 2015.
[44]O. Karadag, U. Kalyoncu, A. Akdogan, Y. Karadag, S. Bilgen, S. Ozbakır et al., Sonographic assessment of carpal tunnel syndrome in rheumatoid arthritis: prevalence and correlation with disease activity, Rheumatology International, vol. 32, no. 8, pp. 2313-2319, 2012.
[45]S. Kaur, S. White, and P. M. Bartold, Periodontal Disease and Rheumatoid Arthritis: A Systematic Review, Journal of Dental Research, vol. 92, no. 5, pp. 399-408, 2013.
[46]J. Ker, L. Wang, J. Rao, and T. Lim, Deep Learning Applications in Medical Image Analysis, IEEE Access, vol. 6, pp. 9375-9389, 2018.
[47]R. Kohavi, A study of cross-validation and bootstrap for accuracy estimation and model selection, presented at the Proceedings of the 14th international joint conference on Artificial intelligence - Volume 2, Montreal, Quebec, Canada, 1995.
[48]A. Krizhevsky, I. Sutskever, and G. E. Hinton, ImageNet classification with deep convolutional neural networks, presented at the Proceedings of the 25th International Conference on Neural Information Processing Systems - Volume 1, Lake Tahoe, Nevada, 2012.
[49]H. M. Kruizenga, M. W. Van Tulder, J. C. Seidell, A. Thijs, H. J. Ader, and M. A. E. Van Bokhorst-de van der Schueren, Effectiveness and cost-effectiveness of early screening and treatment of malnourished patients, The American Journal of Clinical Nutrition, vol. 82, no. 5, pp. 1082-1089, 2005.
[50]C.-C. Kuo, F.-C. Yang, M.-H. Yang, and D.-D. Lee, Predicting the onset of bullous pemphigoid with co-morbidities: A survey based on a nationwide medical database, in 2013 IEEE International Conference on Bioinformatics and Biomedicine, 2013, pp. 30-37.
[51]C. F. Kuo, S. F. Luo, L. C. See, I. J. Chou, H. C. Chang, and K. H. Yu, Rheumatoid arthritis prevalence, incidence, and mortality rates: a nationwide population study in Taiwan, (in English), Rheumatology International, vol. 33, no. 2, pp. 355-360, 2013.
[52]C. Lam, C.-F. Kuan, J. Miser, K.-Y. Hsieh, Y.-A. Fang, Y.-C. Li et al., Emergency department utilization can indicate early diagnosis of digestive tract cancers: A population-based study in Taiwan, Computer Methods and Programs in Biomedicine, vol. 115, no. 3, pp. 103-109, 2014.
[53]G.-C. Lan, C.-H. Lee, Y.-Y. Lee, T. V. S., C.-Y. Chin, M.-L. Day et al., Disease Risk Prediction by Mining Personalized Health Trend Patterns: A Case Study on Diabetes, in Technologies and Applications of Artificial Intelligence (TAAI), 2012 Conference on, 2012, pp. 27-32.
[54]G. V. Lawry, M. L. Finerman, W. N. Hanafee, A. A. Mancuso, P. T. Fan, and R. Bluestone, Laryngeal involvement in rheumatoid arthritis, Arthritis ＆ Rheumatism, vol. 27, no. 8, pp. 873-882, 1984.
[55]C.-H. Lee, J. C.-Y. Chen, and V. S. Tseng, A novel data mining mechanism considering bio-signal and environmental data with applications on asthma monitoring, Computer Methods and Programs in Biomedicine, vol. 101, no. 1, pp. 44-61, 2011.
[56]D. D. Lee and H. S. Seung, Learning the parts of objects by non-negative matrix factorization, Nature, vol. 401, p. 788, 1999.
[57]D. D. Lee and H. S. Seung, Algorithms for non-negative matrix factorization, presented at the Proceedings of the 13th International Conference on Neural Information Processing Systems, Denver, CO, 2000.
[58]L. Y. Lee, M. M. Akhtar, O. Kirresh, and T. Gibson, Interstitial keratitis and sensorineural hearing loss as a manifestation of rheumatoid arthritis: clinical lessons from a rare complication, BMJ Case Reports, vol. 2012, 2012.
[59]N. Lesh, M. J. Zaki, and M. Oglhara, Scalable feature mining for sequential data, IEEE Intelligent Systems and their Applications, vol. 15, no. 2, pp. 48-56, 2000.
[60]J.-N. Liao, T.-F. Chao, C.-J. Liu, K.-L. Wang, S.-J. Chen, T.-C. Tuan et al., Risk and prediction of dementia in patients with atrial fibrillation--a nationwide population-based cohort study, International Journal of Cardiology, vol. 199, pp. 25-30, 2015.
[61]K. P. Liao, T. Cai, V. Gainer, S. Goryachev, Q. Zeng-treitler, S. Raychaudhuri et al., Electronic medical records for discovery research in rheumatoid arthritis, Arthritis care ＆ research, vol. 62, no. 8, pp. 1120-1127, 2010.
[62]R. Liao, Y. Zhang, J. Guan, and S. Zhou, CloudNMF: a MapReduce implementation of nonnegative matrix factorization for large-scale biological datasets, Genomics, proteomics ＆ bioinformatics, vol. 12, no. 1, pp. 48-51, 2014.
[63]C.-C. Lin, C.-I. Li, C.-Y. Hsiao, C.-S. Liu, S.-Y. Yang, C.-C. Lee et al., Time trend analysis of the prevalence and incidence of diagnosed type 2 diabetes among adults in Taiwan from 2000 to 2007: a population-based study, BMC public health, vol. 13, pp. 318-318, 2013.
[64]B. Liu, W. Hsu, and Y. Ma, Integrating classification and association rule mining, presented at the Proceedings of the Fourth International Conference on Knowledge Discovery and Data Mining, New York, NY, 1998.
[65]J. S. Mathias, A. Agrawal, J. Feinglass, A. J. Cooper, D. W. Baker, and A. Choudhary, Development of a 5 year life expectancy index in older adults using predictive mining of electronic health record data, Journal of the American Medical Informatics Association : JAMIA, vol. 20, no. e1, pp. 118-124, 2013.
[66]E. Matsuura, F. Atzeni, P. Sarzi-Puttini, M. Turiel, L. Lopez, and M. Nurmohamed, Is atherosclerosis an autoimmune disease?, BMC Medicine, vol. 12, no. 1, p. 47, 2014.
[67]W. S. McCulloch and W. Pitts, A logical calculus of the ideas immanent in nervous activity, The bulletin of mathematical biophysics, vol. 5, no. 4, pp. 115-133, 1943.
[68]E. McNally, C. Keogh, R. Galvin, and T. Fahey, Diagnostic accuracy of a clinical prediction rule (CPR) for identifying patients with recent-onset undifferentiated arthritis who are at a high risk of developing rheumatoid arthritis: A systematic review and meta-analysis, Seminars in arthritis and rheumatism, vol. 43, no. 4, pp. 498-507, 2014.
[69]E. Mejía-Roa, D. Tabas-Madrid, J. Setoain, C. García, F. Tirado, and A. Pascual-Montano, NMF-mGPU: non-negative matrix factorization on multi-GPU systems, BMC Bioinformatics, journal article vol. 16, no. 1, p. 43, 2015.
[70]R. Miotto, L. Li, B. A. Kidd, and J. T. Dudley, Deep Patient: An Unsupervised Representation to Predict the Future of Patients from the Electronic Health Records, Scientific Reports, vol. 6, p. 26094, 2016.
[71]P. Monsarrat, J.-N. Vergnes, A. Cantagrel, N. Algans, S. Cousty, P. Kemoun et al., Effect of periodontal treatment on the clinical parameters of patients with rheumatoid arthritis: study protocol of the randomized, controlled ESPERA trial, Trials, vol. 14, p. 253, 2013.
[72]C. Mora, J. Díaz, and G. Quintana, Costos directos de la artritis reumatoide temprana en el primer año de atención: simulación de tres situaciones clínicas en un hospital universitario de tercer nivel en Colombia, Biomédica, vol. 29, pp. 43-50, 2009.
[73]T. Ng, L. Chew, and C. W. Yap, A Clinical Decision Support Tool To Predict Survival in Cancer Patients beyond 120 Days after Palliative Chemotherapy, Journal of Palliative Medicine, Article vol. 15, no. 8, pp. 863-869, 2012.
[74]P. Nyirjesy, J. M. Nixon, C. A. Jordan, and H. R. Buckley, Malassezia furfur folliculitis of the vulva: olive oil solves the mystery, Obstetrics and gynecology, vol. 84, no. 4 Pt 2, pp. 710-711, 1994.
[75]P. J. O'Connor, Crystal Deposition Disease and Psoriatic Arthritis, Seminars in musculoskeletal radiology, vol. 17, no. 1, pp. 74-79, 2013.
[76]Y. Ozaki, R. Aoki, T. Kimura, Y. Takashima, and T. Yamada, Characterizing muscular activities using non-negative matrix factorization from EMG channels for driver swings in golf, in 2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2016, pp. 892-895.
[77]P. Paatero and U. Tapper, Positive matrix factorization: A non-negative factor model with optimal utilization of error estimates of data values, Environmetrics, vol. 5, no. 2, pp. 111-126, 1994.
[78]P. Padilla, M. Lopez, J. M. Gorriz, J. Ramirez, D. Salas-Gonzalez, and I. Alvarez, NMF-SVM Based CAD Tool Applied to Functional Brain Images for the Diagnosis of Alzheimer's Disease, IEEE Transactions on Medical Imaging, vol. 31, no. 2, pp. 207-216, 2012.
[79]M. R. L. Paine, J. Kim, R. V. Bennett, R. M. Parry, D. A. Gaul, M. D. Wang et al., Whole Reproductive System Non-Negative Matrix Factorization Mass Spectrometry Imaging of an Early-Stage Ovarian Cancer Mouse Model, PLOS ONE, vol. 11, no. 5, p. e0154837, 2016.
[80]T. Perry, H. Zha, M. E. Oster, P. A. Frias, and M. Braunstein, Utility of a clinical support tool for outpatient evaluation of pediatric chest pain, AMIA ... Annual Symposium proceedings. AMIA Symposium, vol. 2012, pp. 726-733, 2012.
[81]J. R. Quinlan, C4.5: programs for machine learning. Morgan Kaufmann Publishers Inc., 1993, p. 302.
[82]H.-H. Rau, C.-Y. Hsu, Y.-A. Lin, S. Atique, A. Fuad, L.-M. Wei et al., Development of a web-based liver cancer prediction model for type II diabetes patients by using an artificial neural network, Computer Methods and Programs in Biomedicine, vol. 125, pp. 58-65, 2016.
[83]E. Sayers. (2008). E-utilities quick start. Available: http://www.ncbi.nlm.nih.gov/books/NBK25500/
[84]M. Schneider and K. Krüger, Rheumatoid Arthritis—Early Diagnosis and Disease Management, Deutsches Ärzteblatt International, vol. 110, no. 27-28, pp. 477-484, 2013.
[85]D. L. Scott, Early rheumatoid arthritis, British Medical Bulletin, vol. 81-82, no. 1, pp. 97-114, 2007.
[86]N. A. Shadick, N. R. Cook, E. W. Karlson, P. M. Ridker, N. E. Maher, J. E. Manson et al., C-Reactive Protein in the Prediction of Rheumatoid Arthritis in Women, Archives of Internal Medicine, vol. 166, no. 22, pp. 2490-2494, 2006.
[87]G. Shang, A. Richardson, M. E. Gahan, S. Easteal, S. Ohms, and B. A. Lidbury, Predicting the presence of hepatitis B virus surface antigen in Chinese patients by pathology data mining, Journal of Medical Virology, vol. 85, no. 8, pp. 1334-1339, 2013.
[88]C.-C. Shen, L.-Y. Hu, and Y.-H. Hu, Comorbidity study of borderline personality disorder: applying association rule mining to the Taiwan national health insurance research database, BMC medical informatics and decision making, vol. 17, no. 1, pp. 8-8, 2017.
[89]C. Shivade, P. Raghavan, E. Fosler-Lussier, P. J. Embi, N. Elhadad, S. B. Johnson et al., A review of approaches to identifying patient phenotype cohorts using electronic health records, Journal of the American Medical Informatics Association : JAMIA, vol. 21, no. 2, pp. 221-230, 2014.
[90]D. Silver, A. Huang, C. J. Maddison, A. Guez, L. Sifre, G. van den Driessche et al., Mastering the game of Go with deep neural networks and tree search, Nature, Article vol. 529, p. 484, 2016.
[91]J. A. Singh, A. R. Holmgren, and S. Noorbaloochi, Accuracy of veterans administration databases for a diagnosis of rheumatoid arthritis, Arthritis Care ＆ Research, vol. 51, no. 6, pp. 952-957, 2004.
[92]T. L. Skare, R. Nisihara, B. Barbosa, A. da Luz, S. Utiyama, and V. Picceli, Anti-CCP in systemic lupus erythematosus patients: a cross sectional study in Brazilian patients, (in English), Clinical Rheumatology, vol. 32, no. 7, pp. 1065-1070, 2013.
[93]J. S. Smolen, R. Landewé, J. Bijlsma, G. Burmester, K. Chatzidionysiou, M. Dougados et al., EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological disease-modifying antirheumatic drugs: 2016 update, Annals of the Rheumatic Diseases, 2017.
[94]I. Smolik, D. Robinson, and H. S. El-Gabalawy, Periodontitis and rheumatoid arthritis: epidemiologic, clinical, and immunologic associations, Compend Contin Educ Dent, vol. 30, no. 4, pp. 188-90, 192, 194 passim; quiz 198, 210., 2009.
[95]C. Szegedy, S. Ioffe, V. Vanhoucke, and A. A. Alemi, Inception-v4, inception-resnet and the impact of residual connections on learning, in AAAI, 2017, vol. 4, p. 12.
[96]Y. M. Tai and H. W. Chiu, Comorbidity study of ADHD: Applying association rule mining (ARM) to National Health Insurance Database of Taiwan, International Journal of Medical Informatics, vol. 78, no. 12, pp. 75-83, 2009.
[97]V. S. Tseng and C.-H. Lee, Effective temporal data classification by integrating sequential pattern mining and probabilistic induction, Expert Systems with Applications, vol. 36, no. 5, pp. 9524-9532, 2009.
[98]S. A. Turk, D. van Schaardenburg, M. Boers, S. de Boer, C. Fokker, W. F. Lems et al., An unfavorable body composition is common in early arthritis patients: A case control study, PLOS ONE, vol. 13, no. 3, p. e0193377, 2018.
[99]A. H. M. van der Helm-vanMil, S. le Cessie, H. van Dongen, F. C. Breedveld, R. E. M. Toes, and T. W. J. Huizinga, A prediction rule for disease outcome in patients with Recent-onset undifferentiated arthritis: How to guide individual treatment decisions, Arthritis ＆ Rheumatism, vol. 56, no. 2, pp. 433-440, 2007.
[100]V. N. Vapnik, The nature of statistical learning theory. Springer-Verlag New York, Inc., 1995, p. 188.
[101]K. E. Verweij, A. M. E. van Well, J. W. vd Sluijs, and A. Dees, Late Onset Takayasu Arteritis and Rheumatoid Arthritis, Case Reports in Medicine, vol. 2012, p. 523218, 2012, Art. no. 523218.
[102]B. C. M. Wang, P.-N. Hsu, W. Furnback, J. Ney, Y.-W. Yang, C.-H. Fang et al., Estimating the Economic Burden of Rheumatoid Arthritis in Taiwan Using the National Health Insurance Database, Drugs - real world outcomes, vol. 3, no. 1, pp. 107-114, 2016.
[103]W.-Q. Wei, C. Tao, G. Jiang, and C. G. Chute, A high throughput semantic concept frequency based approach for patient identification: a case study using type 2 diabetes mellitus clinical notes, AMIA ... Annual Symposium proceedings. AMIA Symposium, vol. 2010, pp. 857-861, 2010.
[104]W.-Q. Wei, L. A. Bastarache, R. J. Carroll, J. E. Marlo, T. J. Osterman, E. R. Gamazon et al., Evaluating phecodes, clinical classification software, and ICD-9-CM codes for phenome-wide association studies in the electronic health record, PLOS ONE, vol. 12, no. 7, p. e0175508, 2017.
[105]L. Wenmin, H. Jiawei, and P. Jian, CMAR: accurate and efficient classification based on multiple class-association rules, presented at the First IEEE International Conference on Data Mining (ICDM'01), 2001.
[106]H. Xiong, J. Zhang, Y. Huang, K. Leach, and L. E. Barnes, Daehr: A Discriminant Analysis Framework for Electronic Health Record Data and an Application to Early Detection of Mental Health Disorders, ACM Trans. Intell. Syst. Technol., vol. 8, no. 3, pp. 1-21, 2017.
[107]H. Yang, Y.-H. Chen, T.-F. Hsieh, S.-Y. Chuang, and M.-J. Wu, Prediction of Mortality in Incident Hemodialysis Patients: A Validation and Comparison of CHADS2, CHA2DS2, and CCI Scores, PLOS ONE, vol. 11, no. 5, p. e0154627, 2016.
[108]H. Yang and C. Seoighe, Impact of the Choice of Normalization Method on Molecular Cancer Class Discovery Using Nonnegative Matrix Factorization, PLOS ONE, vol. 11, no. 10, p. e0164880, 2016.
[109]X. Yin and J. Han, CPAR: Classification based on Predictive Association Rules, in Proceedings of the 2003 SIAM International Conference on Data Mining, pp. 331-335, 2003.
[110]J. Zhang, H. Xiong, Y. Huang, H. Wu, K. Leach, and L. E. Barnes, M-SEQ: Early detection of anxiety and depression via temporal orders of diagnoses in electronic health data, in 2015 IEEE International Conference on Big Data (Big Data), 2015, pp. 2569-2577.
[111]J. Zhang, J. Gong, and L. Barnes, HCNN: Heterogeneous Convolutional Neural Networks for Comorbid Risk Prediction with Electronic Health Records, in 2017 IEEE/ACM International Conference on Connected Health: Applications, Systems and Engineering Technologies (CHASE), 2017, pp. 214-221.(/font)(/p)