SlideShare a Scribd company logo
DVT assessment using SVM
Daniel Öberg
Introduction
Implementation of a thrombus reporting system
Deep vein thrombosis (DVT) assessment using Support Vector
Machines
Why?
Every year 1.6 per 1000 inhabitants suffer from venous
thrombosis, blood cloths in their veins.
Commonly assessed by a combination of D-Dimer tests and a
questionnaire.
Can we make a system that makes the assessment better?
D-Dimer
A small protein fragment present in the blood after a blood
clot is degraded. High concentration of D-Dimer correlates
with thrombosis, but false positive readings can occur from
liver disease, high rheumatoid factor, inflammation, and many
other factors which makes the sensitivity low.
Wells Score
Gold standard for DVT clinical assessment, named Wells Score.
A simple scoring system and a yes or no questionnaire
regarding the patients medical history.
Very quick assessment as it only has 9 significant variables.
Assessment in real world
A combination of Wells Score and a D-dimer test reliably
excludes DVT
Often no need for painful imaging studies such as compression
ultrasonography
Compression Ultrasonography
Compressing the veins
Send pulses of ultrasound into the leg to find material with
different density
Note absence or presence of occluded veins
Needed to be reported
Compression Ultrasonography
Figure 1:Compression Ultrasonography
Deep Vein Thrombosis
Blood is constantly coagulating and dissolving.
This balance between the stimulating and inhibitory is sensitive.
Coagulated blood could produce blood clots
Dangerous internal bleeding otherwise
Rate of occurrence
If part of a cloth breaks free and is allowed to travel to the lung , we
call it a pulmonary embolism . A serious condition that 1000
patients a year die from just in Sweden. These figures should be
compared to the ~4000 who gets diagnosed with pulmonary
embolism or ~8000 patients who gets diagnosed with deep vein
thrombosis.
Costs
The Swedish hospitals costs of venous thromboembolism alone
was estimated to 0.375 billion SEK in 1999.
Difficulty to confirm diagnosis without thrombosonography
Pretest
The decision to order thrombosonography is, in several
guidelines, in large part done based entirely on pretest risk
assessment like Wells Score . The patients with low risk get
D-dimer blood test and only go on to ultrasonography if the
test is positive. The ones with high risk goes straight to
ultrasonography without getting a D-dimer test.
Technical Background
Can we improve Wells Score?
With machine learning algorithms we can figure out how to
assess deep vein thrombosis by generalising from examples.
As more data would become available in the reporting system
the better the assessment would become.
Support Vector Machines
Developed by Cortes & Vapnik for classification.
Finding the maximal margin of separation between two classes,
which can be seen as the generalisation, for linearly separable
patterns.
Support Vector Machines
With support vector classifiers the hyperplane can be described by
the unknown, u, and the margin-vector, w (see figure 2):
w · u ≥ C (1)
Support Vector Machines
Figure 2:Optimal hyperplane
Support Vector Machines
We are interested in which side the u vector is in so we project the
u onto w and if this is bigger than some constant C then we say
that u a positive sample.
Support Vector Machines
By setting b = −C we get our decision rule:
(w · u) + b = 0, w ∈ RN
, u ∈ RN
, b ∈ R (2)
which corresponds to the decision function:
f (x) = sign((w · u) + b) (3)
Support Vector Machines
The problem is that we do not know neither the w nor the b.
However adding additional constraints we can calculate them.
Taking a positive and negative sample and arbitrary setting it to
bigger and smaller than one respectively:
w · x+ ≥ 1 (4)
w · x− ≤ 1 (5)
Support Vector Machines
Then introducing a variable yi that is +1 for positive samples and
−1 for negative samples we can combine these into:
yi ∗ (xi · w + b) ≥ 1 (6)
Support Vector Machines
And with that we can add the extra constraint that:
yi ∗ (xi · w + b) − 1 = 0 (7)
should be were xi is on the margin.
Support Vector Machines
Now if we want the widest margin possible we could take the
difference of a negative and positive sample on the margins and
project it onto a unit normal.
WIDTH = (x+ − x−) ·
w
w
(8)
Support Vector Machines
Which can be simplified further to
(x+ − x−) ·
w
w
=
2
w
(9)
Support Vector Machines
The way one maximize this in the support vector machines
algorithm is to use Lagrange multipliers.
Radial Basis Function
But with the so called kernel-trick , were we map data into a richer
feature space then construct a hyperplane in that space, we are able
to classify points that were not linearly separable in its previous
space.
We call the function that maps from the vector x to another input
space φ(x).
Radial Basis Function
By doing this simple transformation we know need to maximize:
K(x, y) = φ(x) · φ(y) (10)
Radial Basis Function
Don’t need φ(x) on its own but can instead focus on K(x, y) which
we call our kernel function.
By using the radial basis kernel (RBF):
K(x, y) = e−γ x−y 2
(11)
where γ is a chosen constant, we get a great and fast nonlinear
kernel.
Environment
Apple iOS
C++ and Objective-C
OpenCV developed by Intel Russia research center in Nizhny
Novgorod for realtime computer vision. This library contains
implementations for both RBF and linear kernels as it adopted
the SVM/C++ library libsvm.
Preprocessing
159 anonymous patients DVT journals
From the journals we extracted the Wells score information and
whether a DVT or occlusion were found.
Table 1:Count of each label in dataset
DVT Nothing found
33 126
Already gone through a Wells score, heavy bias.
Yes and no converted to 1.0 and -1.0 respectively.
Training
Optimized C and gamma values
5 folds
C-values between 2−5 and 215
Gamma-values between 2−15 and 23
Baseline
Wells Score by Philip S. Wells
Gold standard for DVT assessment.
Baseline
Table 2:Wells score
Variable
Points
Cancer treatment during the past 6 months +1
Lower leg paralysis or plastering +1
Bed rest > 3 days or surgery < 4 weeks +1
Pain on palpation of deep veins +1
Swelling of entire leg +1
Diameter difference on affected calf > 3 cm +1
Pitting oedema (affected side only) +1
Dilated superficial veins (affected side) +1
Alternative diagnosis at least as probable as DVT -2
Baseline
Table 3:Clinical probability for Wells score
Low 0 total
Intermediate 1-2 total
High > 2 total
Results
Figure 3:Assessment Model
Results
Figure 4:Assessment Statistics
Results
Figure 5:Features
Results
Figure 6:Feature Statistics
Results
Figure 7:Journal
Results
Figure 8:Journal Questions
Results
Figure 9:Sonography
Results
Figure 10:Linear SVM with soft margins and different error costs compared
to Wells Score. C : 12.5
Results
Figure 11:SVM with RBF kernel compared to Wells Score.
RoC
Figure 12:
Results
Class Weight Accuracy
SVM RBF 0.9226 58.49%
Wells Score - MEDIUM N/A 23.12%
SVM RBF 0.8196 81.11%
SVM Linear 0.9193 65.40%
Wells Score - HIGH N/A 58.49%
Results
Class Weight Sensitivity Specificity
SVM RBF 0.9226 100.00% 35.71%
Wells Score - MEDIUM N/A 97.05% 3.17%
SVM RBF 0.8196 66.66% 84.92%
SVM Linear 0.9193 63.63% 65.87%
Wells Score - HIGH N/A 60.60% 57.93%
Balanced Error Rate
BER =
FP/(TN + FP) + FN/(FN + TP)
2
Average of both the error rate of the positive class and the error
rate of the negative class.
Diagnostic Odds Ratio
DOR =
TP/FP
FN/TN
Ratio of the odds of the test being positive if the subject has a
disease relative to the odds of the test being positive if the subject
does not have the disease.
Results
Class Weight BCR DOR
SVM RBF 0.9226 67.85% ∞
Wells Score - MEDIUM N/A 50.11% 01.08
SVM RBF 0.8196 75.79% 11.26
SVM Linear 0.9193 64.75% 03.37
Wells Score - HIGH N/A 59.27% 02.11
Conclusion
Possible to improve DVT assessment by using SVMs.
Able to get 100% sensitivity with 58% accuracy.
Balanced Classification Rate for Well Score was at most 02.11
whilst our highest benchmarked was 11.26 .
Using different error costs (DEC) we can tweak the sensitivity
and specificity.

More Related Content

Viewers also liked

Social media: cosa dovrebbero sapere le aziende italiane
Social media: cosa dovrebbero sapere le aziende italianeSocial media: cosa dovrebbero sapere le aziende italiane
Social media: cosa dovrebbero sapere le aziende italiane
Nicoletta Boldrini
 
CERN and Japan
CERN and JapanCERN and Japan
CERN and Japan
Masahiro Kuze
 
Historia de-la-agricultura
Historia de-la-agriculturaHistoria de-la-agricultura
Historia de-la-agricultura
CinthyaSamanthaRM
 
Melisa rodriguez5b
Melisa rodriguez5bMelisa rodriguez5b
Melisa rodriguez5b
meroara
 
Branding fjsc (1)
Branding fjsc (1)Branding fjsc (1)
Branding fjsc (1)
Ghencea Andreea
 
Hoe slachtoffer bouwen reddingsopdracht
Hoe slachtoffer bouwen reddingsopdrachtHoe slachtoffer bouwen reddingsopdracht
Hoe slachtoffer bouwen reddingsopdracht
RoboCupJuniorBelgium
 
Rassegna della normativa sui tirocini in Sicilia
Rassegna della normativa sui tirocini in SiciliaRassegna della normativa sui tirocini in Sicilia
Rassegna della normativa sui tirocini in Sicilia
Giovanni Nocera
 
Tập huấn phần mềm Công cụ dành cho GV THCS, THPT
Tập huấn phần mềm Công cụ dành cho GV THCS, THPTTập huấn phần mềm Công cụ dành cho GV THCS, THPT
Tập huấn phần mềm Công cụ dành cho GV THCS, THPT
Bùi Việt Hà
 
Brand & Design on YouTube Channels
Brand & Design on YouTube ChannelsBrand & Design on YouTube Channels
Brand & Design on YouTube Channels
BOB Liuzzo
 
ética jurídica
ética jurídicaética jurídica
ética jurídica
CinthyaSamanthaRM
 
Buduschee bez-terrorizma
Buduschee bez-terrorizmaBuduschee bez-terrorizma
Sponsopship
SponsopshipSponsopship
Sponsopship
Alexamaribel21
 
Harry Potter
Harry PotterHarry Potter
Harry Potter
read in europe
 
Ege 2015 licey66
Ege 2015 licey66Ege 2015 licey66
Blis insights: the influence of location on cross-screen advertising
Blis insights: the influence of location on cross-screen advertisingBlis insights: the influence of location on cross-screen advertising
Blis insights: the influence of location on cross-screen advertising
Blis
 

Viewers also liked (15)

Social media: cosa dovrebbero sapere le aziende italiane
Social media: cosa dovrebbero sapere le aziende italianeSocial media: cosa dovrebbero sapere le aziende italiane
Social media: cosa dovrebbero sapere le aziende italiane
 
CERN and Japan
CERN and JapanCERN and Japan
CERN and Japan
 
Historia de-la-agricultura
Historia de-la-agriculturaHistoria de-la-agricultura
Historia de-la-agricultura
 
Melisa rodriguez5b
Melisa rodriguez5bMelisa rodriguez5b
Melisa rodriguez5b
 
Branding fjsc (1)
Branding fjsc (1)Branding fjsc (1)
Branding fjsc (1)
 
Hoe slachtoffer bouwen reddingsopdracht
Hoe slachtoffer bouwen reddingsopdrachtHoe slachtoffer bouwen reddingsopdracht
Hoe slachtoffer bouwen reddingsopdracht
 
Rassegna della normativa sui tirocini in Sicilia
Rassegna della normativa sui tirocini in SiciliaRassegna della normativa sui tirocini in Sicilia
Rassegna della normativa sui tirocini in Sicilia
 
Tập huấn phần mềm Công cụ dành cho GV THCS, THPT
Tập huấn phần mềm Công cụ dành cho GV THCS, THPTTập huấn phần mềm Công cụ dành cho GV THCS, THPT
Tập huấn phần mềm Công cụ dành cho GV THCS, THPT
 
Brand & Design on YouTube Channels
Brand & Design on YouTube ChannelsBrand & Design on YouTube Channels
Brand & Design on YouTube Channels
 
ética jurídica
ética jurídicaética jurídica
ética jurídica
 
Buduschee bez-terrorizma
Buduschee bez-terrorizmaBuduschee bez-terrorizma
Buduschee bez-terrorizma
 
Sponsopship
SponsopshipSponsopship
Sponsopship
 
Harry Potter
Harry PotterHarry Potter
Harry Potter
 
Ege 2015 licey66
Ege 2015 licey66Ege 2015 licey66
Ege 2015 licey66
 
Blis insights: the influence of location on cross-screen advertising
Blis insights: the influence of location on cross-screen advertisingBlis insights: the influence of location on cross-screen advertising
Blis insights: the influence of location on cross-screen advertising
 

Similar to slides

Conference_paper.pdf
Conference_paper.pdfConference_paper.pdf
Conference_paper.pdf
NarenRajVivek
 
Pm m23 & pmnm06 week 3 lectures 2015
Pm m23 & pmnm06 week 3 lectures 2015Pm m23 & pmnm06 week 3 lectures 2015
Pm m23 & pmnm06 week 3 lectures 2015
pdiddyboy2
 
Healthcare deserts: How accessible is US healthcare?
Healthcare deserts: How accessible is US healthcare?Healthcare deserts: How accessible is US healthcare?
Healthcare deserts: How accessible is US healthcare?
Data Con LA
 
IRJET- Diagnosis of Diabetic Retinopathy using Machine Learning Algorithms
IRJET- Diagnosis of Diabetic Retinopathy using Machine Learning AlgorithmsIRJET- Diagnosis of Diabetic Retinopathy using Machine Learning Algorithms
IRJET- Diagnosis of Diabetic Retinopathy using Machine Learning Algorithms
IRJET Journal
 
Presentation
PresentationPresentation
Presentation
Weston Eldredge
 
A Multi-Objective Genetic Algorithm for Pruning Support Vector Machines
A Multi-Objective Genetic Algorithm for Pruning Support Vector MachinesA Multi-Objective Genetic Algorithm for Pruning Support Vector Machines
A Multi-Objective Genetic Algorithm for Pruning Support Vector Machines
Mohamed Farouk
 
Unknown power power point unknown power point
Unknown power power point unknown power pointUnknown power power point unknown power point
Unknown power power point unknown power point
xmendquick
 
bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...
bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...
bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...
Matt Moores
 
Abdominal Bleeding Region Detection in Wireless Capsule Endoscopy Videos Usin...
Abdominal Bleeding Region Detection in Wireless Capsule Endoscopy Videos Usin...Abdominal Bleeding Region Detection in Wireless Capsule Endoscopy Videos Usin...
Abdominal Bleeding Region Detection in Wireless Capsule Endoscopy Videos Usin...
journal ijrtem
 
Breast cancer diagnosis and recurrence prediction using machine learning tech...
Breast cancer diagnosis and recurrence prediction using machine learning tech...Breast cancer diagnosis and recurrence prediction using machine learning tech...
Breast cancer diagnosis and recurrence prediction using machine learning tech...
eSAT Journals
 
Encoder for (7,3) cyclic code using matlab
Encoder for (7,3) cyclic code using matlabEncoder for (7,3) cyclic code using matlab
Encoder for (7,3) cyclic code using matlab
SneheshDutta
 
Digitech Ophthalmologist (Techlead Image Processing)
Digitech Ophthalmologist (Techlead Image Processing)Digitech Ophthalmologist (Techlead Image Processing)
Digitech Ophthalmologist (Techlead Image Processing)
techlead-india
 
Lec.5 - Analysis of Measurement Data.pptx
Lec.5 - Analysis of Measurement Data.pptxLec.5 - Analysis of Measurement Data.pptx
Lec.5 - Analysis of Measurement Data.pptx
Shimaa Barakat
 
Teleradiology Overview Systems and Applications - Sanjoy Sanyal
Teleradiology Overview Systems and Applications - Sanjoy SanyalTeleradiology Overview Systems and Applications - Sanjoy Sanyal
Teleradiology Overview Systems and Applications - Sanjoy Sanyal
Sanjoy Sanyal
 
Cloud-based ECG classification with mobile interface.pptx
Cloud-based ECG classification with mobile interface.pptxCloud-based ECG classification with mobile interface.pptx
Cloud-based ECG classification with mobile interface.pptx
Shamman Noor Shoudha
 
SVM - Functional Verification
SVM - Functional VerificationSVM - Functional Verification
SVM - Functional Verification
Sai Kiran Kadam
 
(short presentation) A Wearable Monitoring System for Continuously Assessing ...
(short presentation) A Wearable Monitoring System for Continuously Assessing ...(short presentation) A Wearable Monitoring System for Continuously Assessing ...
(short presentation) A Wearable Monitoring System for Continuously Assessing ...
Meghan Hegarty
 
A Wearable Monitoring System for Continuously Assessing the Health of the Per...
A Wearable Monitoring System for Continuously Assessing the Health of the Per...A Wearable Monitoring System for Continuously Assessing the Health of the Per...
A Wearable Monitoring System for Continuously Assessing the Health of the Per...
Meghan Hegarty
 
Interpretation of electrocardiography (ECG) by using polynomial function simu...
Interpretation of electrocardiography (ECG) by using polynomial function simu...Interpretation of electrocardiography (ECG) by using polynomial function simu...
Interpretation of electrocardiography (ECG) by using polynomial function simu...
mustafatacettin1
 
P1121133727
P1121133727P1121133727
P1121133727
Ashraf Aboshosha
 

Similar to slides (20)

Conference_paper.pdf
Conference_paper.pdfConference_paper.pdf
Conference_paper.pdf
 
Pm m23 & pmnm06 week 3 lectures 2015
Pm m23 & pmnm06 week 3 lectures 2015Pm m23 & pmnm06 week 3 lectures 2015
Pm m23 & pmnm06 week 3 lectures 2015
 
Healthcare deserts: How accessible is US healthcare?
Healthcare deserts: How accessible is US healthcare?Healthcare deserts: How accessible is US healthcare?
Healthcare deserts: How accessible is US healthcare?
 
IRJET- Diagnosis of Diabetic Retinopathy using Machine Learning Algorithms
IRJET- Diagnosis of Diabetic Retinopathy using Machine Learning AlgorithmsIRJET- Diagnosis of Diabetic Retinopathy using Machine Learning Algorithms
IRJET- Diagnosis of Diabetic Retinopathy using Machine Learning Algorithms
 
Presentation
PresentationPresentation
Presentation
 
A Multi-Objective Genetic Algorithm for Pruning Support Vector Machines
A Multi-Objective Genetic Algorithm for Pruning Support Vector MachinesA Multi-Objective Genetic Algorithm for Pruning Support Vector Machines
A Multi-Objective Genetic Algorithm for Pruning Support Vector Machines
 
Unknown power power point unknown power point
Unknown power power point unknown power pointUnknown power power point unknown power point
Unknown power power point unknown power point
 
bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...
bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...
bayesImageS: Bayesian computation for medical Image Segmentation using a hidd...
 
Abdominal Bleeding Region Detection in Wireless Capsule Endoscopy Videos Usin...
Abdominal Bleeding Region Detection in Wireless Capsule Endoscopy Videos Usin...Abdominal Bleeding Region Detection in Wireless Capsule Endoscopy Videos Usin...
Abdominal Bleeding Region Detection in Wireless Capsule Endoscopy Videos Usin...
 
Breast cancer diagnosis and recurrence prediction using machine learning tech...
Breast cancer diagnosis and recurrence prediction using machine learning tech...Breast cancer diagnosis and recurrence prediction using machine learning tech...
Breast cancer diagnosis and recurrence prediction using machine learning tech...
 
Encoder for (7,3) cyclic code using matlab
Encoder for (7,3) cyclic code using matlabEncoder for (7,3) cyclic code using matlab
Encoder for (7,3) cyclic code using matlab
 
Digitech Ophthalmologist (Techlead Image Processing)
Digitech Ophthalmologist (Techlead Image Processing)Digitech Ophthalmologist (Techlead Image Processing)
Digitech Ophthalmologist (Techlead Image Processing)
 
Lec.5 - Analysis of Measurement Data.pptx
Lec.5 - Analysis of Measurement Data.pptxLec.5 - Analysis of Measurement Data.pptx
Lec.5 - Analysis of Measurement Data.pptx
 
Teleradiology Overview Systems and Applications - Sanjoy Sanyal
Teleradiology Overview Systems and Applications - Sanjoy SanyalTeleradiology Overview Systems and Applications - Sanjoy Sanyal
Teleradiology Overview Systems and Applications - Sanjoy Sanyal
 
Cloud-based ECG classification with mobile interface.pptx
Cloud-based ECG classification with mobile interface.pptxCloud-based ECG classification with mobile interface.pptx
Cloud-based ECG classification with mobile interface.pptx
 
SVM - Functional Verification
SVM - Functional VerificationSVM - Functional Verification
SVM - Functional Verification
 
(short presentation) A Wearable Monitoring System for Continuously Assessing ...
(short presentation) A Wearable Monitoring System for Continuously Assessing ...(short presentation) A Wearable Monitoring System for Continuously Assessing ...
(short presentation) A Wearable Monitoring System for Continuously Assessing ...
 
A Wearable Monitoring System for Continuously Assessing the Health of the Per...
A Wearable Monitoring System for Continuously Assessing the Health of the Per...A Wearable Monitoring System for Continuously Assessing the Health of the Per...
A Wearable Monitoring System for Continuously Assessing the Health of the Per...
 
Interpretation of electrocardiography (ECG) by using polynomial function simu...
Interpretation of electrocardiography (ECG) by using polynomial function simu...Interpretation of electrocardiography (ECG) by using polynomial function simu...
Interpretation of electrocardiography (ECG) by using polynomial function simu...
 
P1121133727
P1121133727P1121133727
P1121133727
 

slides

  • 1. DVT assessment using SVM Daniel Öberg
  • 2. Introduction Implementation of a thrombus reporting system Deep vein thrombosis (DVT) assessment using Support Vector Machines
  • 3. Why? Every year 1.6 per 1000 inhabitants suffer from venous thrombosis, blood cloths in their veins. Commonly assessed by a combination of D-Dimer tests and a questionnaire. Can we make a system that makes the assessment better?
  • 4. D-Dimer A small protein fragment present in the blood after a blood clot is degraded. High concentration of D-Dimer correlates with thrombosis, but false positive readings can occur from liver disease, high rheumatoid factor, inflammation, and many other factors which makes the sensitivity low.
  • 5. Wells Score Gold standard for DVT clinical assessment, named Wells Score. A simple scoring system and a yes or no questionnaire regarding the patients medical history. Very quick assessment as it only has 9 significant variables.
  • 6. Assessment in real world A combination of Wells Score and a D-dimer test reliably excludes DVT Often no need for painful imaging studies such as compression ultrasonography
  • 7. Compression Ultrasonography Compressing the veins Send pulses of ultrasound into the leg to find material with different density Note absence or presence of occluded veins Needed to be reported
  • 9. Deep Vein Thrombosis Blood is constantly coagulating and dissolving. This balance between the stimulating and inhibitory is sensitive. Coagulated blood could produce blood clots Dangerous internal bleeding otherwise
  • 10. Rate of occurrence If part of a cloth breaks free and is allowed to travel to the lung , we call it a pulmonary embolism . A serious condition that 1000 patients a year die from just in Sweden. These figures should be compared to the ~4000 who gets diagnosed with pulmonary embolism or ~8000 patients who gets diagnosed with deep vein thrombosis.
  • 11. Costs The Swedish hospitals costs of venous thromboembolism alone was estimated to 0.375 billion SEK in 1999. Difficulty to confirm diagnosis without thrombosonography
  • 12. Pretest The decision to order thrombosonography is, in several guidelines, in large part done based entirely on pretest risk assessment like Wells Score . The patients with low risk get D-dimer blood test and only go on to ultrasonography if the test is positive. The ones with high risk goes straight to ultrasonography without getting a D-dimer test.
  • 14. Can we improve Wells Score? With machine learning algorithms we can figure out how to assess deep vein thrombosis by generalising from examples. As more data would become available in the reporting system the better the assessment would become.
  • 15. Support Vector Machines Developed by Cortes & Vapnik for classification. Finding the maximal margin of separation between two classes, which can be seen as the generalisation, for linearly separable patterns.
  • 16. Support Vector Machines With support vector classifiers the hyperplane can be described by the unknown, u, and the margin-vector, w (see figure 2): w · u ≥ C (1)
  • 17. Support Vector Machines Figure 2:Optimal hyperplane
  • 18. Support Vector Machines We are interested in which side the u vector is in so we project the u onto w and if this is bigger than some constant C then we say that u a positive sample.
  • 19. Support Vector Machines By setting b = −C we get our decision rule: (w · u) + b = 0, w ∈ RN , u ∈ RN , b ∈ R (2) which corresponds to the decision function: f (x) = sign((w · u) + b) (3)
  • 20. Support Vector Machines The problem is that we do not know neither the w nor the b. However adding additional constraints we can calculate them. Taking a positive and negative sample and arbitrary setting it to bigger and smaller than one respectively: w · x+ ≥ 1 (4) w · x− ≤ 1 (5)
  • 21. Support Vector Machines Then introducing a variable yi that is +1 for positive samples and −1 for negative samples we can combine these into: yi ∗ (xi · w + b) ≥ 1 (6)
  • 22. Support Vector Machines And with that we can add the extra constraint that: yi ∗ (xi · w + b) − 1 = 0 (7) should be were xi is on the margin.
  • 23. Support Vector Machines Now if we want the widest margin possible we could take the difference of a negative and positive sample on the margins and project it onto a unit normal. WIDTH = (x+ − x−) · w w (8)
  • 24. Support Vector Machines Which can be simplified further to (x+ − x−) · w w = 2 w (9)
  • 25. Support Vector Machines The way one maximize this in the support vector machines algorithm is to use Lagrange multipliers.
  • 26. Radial Basis Function But with the so called kernel-trick , were we map data into a richer feature space then construct a hyperplane in that space, we are able to classify points that were not linearly separable in its previous space. We call the function that maps from the vector x to another input space φ(x).
  • 27. Radial Basis Function By doing this simple transformation we know need to maximize: K(x, y) = φ(x) · φ(y) (10)
  • 28. Radial Basis Function Don’t need φ(x) on its own but can instead focus on K(x, y) which we call our kernel function. By using the radial basis kernel (RBF): K(x, y) = e−γ x−y 2 (11) where γ is a chosen constant, we get a great and fast nonlinear kernel.
  • 29. Environment Apple iOS C++ and Objective-C OpenCV developed by Intel Russia research center in Nizhny Novgorod for realtime computer vision. This library contains implementations for both RBF and linear kernels as it adopted the SVM/C++ library libsvm.
  • 30. Preprocessing 159 anonymous patients DVT journals From the journals we extracted the Wells score information and whether a DVT or occlusion were found. Table 1:Count of each label in dataset DVT Nothing found 33 126 Already gone through a Wells score, heavy bias. Yes and no converted to 1.0 and -1.0 respectively.
  • 31. Training Optimized C and gamma values 5 folds C-values between 2−5 and 215 Gamma-values between 2−15 and 23
  • 32. Baseline Wells Score by Philip S. Wells Gold standard for DVT assessment.
  • 33. Baseline Table 2:Wells score Variable Points Cancer treatment during the past 6 months +1 Lower leg paralysis or plastering +1 Bed rest > 3 days or surgery < 4 weeks +1 Pain on palpation of deep veins +1 Swelling of entire leg +1 Diameter difference on affected calf > 3 cm +1 Pitting oedema (affected side only) +1 Dilated superficial veins (affected side) +1 Alternative diagnosis at least as probable as DVT -2
  • 34. Baseline Table 3:Clinical probability for Wells score Low 0 total Intermediate 1-2 total High > 2 total
  • 42. Results Figure 10:Linear SVM with soft margins and different error costs compared to Wells Score. C : 12.5
  • 43. Results Figure 11:SVM with RBF kernel compared to Wells Score.
  • 45. Results Class Weight Accuracy SVM RBF 0.9226 58.49% Wells Score - MEDIUM N/A 23.12% SVM RBF 0.8196 81.11% SVM Linear 0.9193 65.40% Wells Score - HIGH N/A 58.49%
  • 46. Results Class Weight Sensitivity Specificity SVM RBF 0.9226 100.00% 35.71% Wells Score - MEDIUM N/A 97.05% 3.17% SVM RBF 0.8196 66.66% 84.92% SVM Linear 0.9193 63.63% 65.87% Wells Score - HIGH N/A 60.60% 57.93%
  • 47. Balanced Error Rate BER = FP/(TN + FP) + FN/(FN + TP) 2 Average of both the error rate of the positive class and the error rate of the negative class.
  • 48. Diagnostic Odds Ratio DOR = TP/FP FN/TN Ratio of the odds of the test being positive if the subject has a disease relative to the odds of the test being positive if the subject does not have the disease.
  • 49. Results Class Weight BCR DOR SVM RBF 0.9226 67.85% ∞ Wells Score - MEDIUM N/A 50.11% 01.08 SVM RBF 0.8196 75.79% 11.26 SVM Linear 0.9193 64.75% 03.37 Wells Score - HIGH N/A 59.27% 02.11
  • 50. Conclusion Possible to improve DVT assessment by using SVMs. Able to get 100% sensitivity with 58% accuracy. Balanced Classification Rate for Well Score was at most 02.11 whilst our highest benchmarked was 11.26 . Using different error costs (DEC) we can tweak the sensitivity and specificity.