Xây Dựng Hệ Thống Hỗ Trợ Hẩn Đoán Bệnh Sử Dụng Neural Networks.pdf

        2019 Tai n[.]

 1

 2

 3

 4

 5

 7

1.1. 7

1.2. 9

1.3. 9

1.4. 9

1.5. 10

 11

2.1.  11

2.2.1  11

2.2.2  11

2.2.3  12

2.2.4  13

2.3. - 15

2.3.1  15

2.3.2  16

2.3.3  19

2.3.4  20

  25

3.1. 25

3.2. 26

3.3. 28

3.4. 30

3.4.1  30

3.4.2  31

3.5. - 33

3.5.1  33

3.5.2  39

3.5.3  41

3.5.4 -ron 45

3.5.5 -ron 47

 52

4.1. 52

4.2. 52

4.3. 54

4.4. 

 55

4.4.1  55

4.4.2  56

4.4.3  - 57

4.5. 58

 61

5.1. 61

5.2. 61

 62

FPG Fact Precedence Graph 

UCI Center for Machine Learning

and Intelligent Systems at the University of California, Irvine





lifornia, Irvine HbA1c Glycated hemoglobin 

glucose

4



 13

 18

 29

 31

 32

 32

 33

  34

 35

 35

 36

 38

 40

-ron 42

 overfitting sau cùng 4000 chu kì epoch 51

 53

- 54

 - 55     56

  58

5



Hình 1:  11

 Fact Precedence Graph (FPG) 14

-ron 16

- 19

- 20

 21

 22

 23

-ron 23

Hình 10:  25

- 27

 30

 40

  -  41

  -  42

 44

 44

-ron 45

  - 46

 47  - 48

 dropout sau 500 chu kì epoch 49

 L2 regularization sau 500 chu kì epoch 50

 51

6

- 52

  53

   Blood Pressure và Skin Thickness 57

 59

H 60

 60

       4.77% Trong mô hình

 g -ron              

       nh     -ron 

11

  : 2.1 H   

12



 con 

Mã lut u ki n  K t lun  tin c y (F) 

Biu di n: 

N u (s ki n(  n( ) ^ s ki n(  )) thì  ((s ki n( ))  (y%)

Fact Precedence Graph (FPG), là

    th th hi n m i liên quan gi a s ki n này v i s ki n khác       

20

 ron 

Hình 9 Công th c tính output cho 1 node trong m ng -:    ron

Vi gii thu t lan truy c s  d ng là hàm Sigmoid

       Gradient Descent   

ng s t i thi u c a hàm l i T ng giá tr l       c tính theo công thc:

24

 

g s b ng vi  o hàm c a E trên t ng tr ng s    

V i m ng -   ron o hàm c a E trên tr ng s    sau:

 w5 = w5 5 





  uá trìn    -ron      

 

Blood Pressure Number  0

Skin Thickness Number  0

31

trong :

  ,    

v kho ng [0,1] b  MinMaxScaling theo công thc sau:



 cacbohydrat,  và protein khi hoóc môn insulin   

  luôn cao; trong





0

SK10         



Diabetes Pedigree Function

90mm

Skin Thickness 2

37

SK16  Skin Thickness 4 SK17        

Final Conclusions Output UnitsSupporting Facts Input Units Intermediate Conclusions Hidden Units Dependencies Weighted Connections

 trung gian - Intermediate Conclusions

45

 -roncho các node       và các node khác  

    -ron       

sau:

Hình 18: Xây -ron3.5.4  -ron

và 



          

 -ron

 



                (Rectified linear unit):

 Hàm truy gin a l p n vào lu ra là hàm Sigmoid:

Early stopping + L2 regularization 86.36%

Early stopping + L2 regularization + Dropout 81.81%

 [5] K c



 Máy Dell E6440

 CPU Core i7 - 4600M 2 core 4 thread

 GPU AMD Radeon HD 8690M



Hình 25-ron 

55

 18 -ron

Probabilistic neural network (PNN) [16] 81.49%

Small-world neural network (SW-FFANN)[17] 93.06%



4.4.3   -ron 

               

là Skin Thickness và Blood Pressure ,  Sau

- 



[3] Y Chen, J Elenee Argent      

Cognitive Computing Can Be Applied to Big Data Challenges in Life

 Clinical Therapeutics, vol 38, no 4, pp 688 701, Apr 2016

[4] Q K Al-      

International Journal of Computer Science Issues, vol 8, no 2, p 6, 2011

[5] -indians-

[6] 

[7] Phan Huy Khánh, Giáo Trình Hệ Chuyên Gia 2004

[8] K Gurney, An Introduction to Neural Networks University of Sheffield, 1997

[9]      

         Engineering Research, 2015, vol 5

[14]        -based artificial neural

Artificial Intelligence, vol 70, no 12, pp 119165, Oct 1994

[15] N Srivastava, G Hinton, A Krizhevsky, I Sutskever, and R

Salak        



[18]         -layer Feed Forward

Neural Network Approach for D   2018 11th

International Conference on Developments in eSystems Engineering

of Diabetes Using Ar     Engineering

Vibration, Communication and Information Processing, vol 478, K Ray, S

N Sharan, S Rawat, S K Jain, S Srivastava, and A Bandyopadhyay, Eds Singapore: The International Conference On Engineering Vibration Communication and Information Processing, 2018, pp 679 687 

[23] 

       Circuits, System and Simulation, 2011

[24] -Based Artificial Neural Network

       

Department of Electrical Engineering and Computer Science, p 7, 2008

[25] B Strack et al. of HbA1c Measurement on Hospital Readmission

BioMed research international, 2014, vol 2014

D Peter, S L Fernandes, C Eduardo Thomaz, and S Viriri, Eds Cham: Springer International Publishing, 2019, pp 179 187 

[30] C C Geddes, J G Fox, M E Allison, J M Boulton-Jones, and K

Simpdeveloping progressive IgA nephropathy more accurately than

  Nephrology, dialysis, transplantation : official publication of the European Dialysis and Transplant Association -

[31]          



[32] A G Karegowda, A S Manjunath, and M A Jayaram, Application of

Genetic Algorithm Optimized Neural Network Connection Weights for Medical Diagnosis of Pima Indians Diabetes 2011

[33]         

Predicting Early Diabetes Patient Hospital Readmittance to Help Optimize

Leland Stanford Junior University, Dec 2017

[34]       -Day Hospital

      

International Journal of Advanced Computer Science and Applications,

vol 10, no 2, 2019

65

[35]           

     Expert Systems with

Ojha, Eds Singapore: Springer Singapore, 2019, pp 1123

[38] ary correlated gravitational

search algorithm (ECGS) with genetic optimized Hopfield neural network (GHNN)         

[39] J S Majeed Alneamy, Z A Hameed Alnaish, S Z Mohd Hashim, and R

        networks with a teaching learning-based optimization algorithm for

Computers in Biology and Medicine, vol 112,

p 103348, Sep 2019