Machine Learning in Medicine: Literature Review

This literature review paper provides an in-depth analysis of machine learning in medicine. The paper is divided into five sections, including a broad and focused scan, research problem, background details, literature review, and summary and discussion.

qwertyuiopasdfghjklzxcvbnmqwe
rtyuiopasdfghjklzxcvbnmqwertyu
iopasdfghjklzxcvbnmqwertyuiopa
sdfghjklzxcvbnmqwertyuiopasdfg
hjklzxcvbnmqwertyuiopasdfghjkl
zxcvbnmqwertyuiopasdfghjklzxcv
bnmqwertyuiopasdfghjklzxcvbnm
qwertyuiopasdfghjklzxcvbnmqwe
rtyuiopasdfghjklzxcvbnmqwertyu
iopasdfghjklzxcvbnmqwertyuiopa
sdfghjklzxcvbnmqwertyuiopasdfg
hjklzxcvbnmqwertyuiopasdfghjkl
zxcvbnmqwertyuiopasdfghjklzxcv
Introduction to Research
Assignment 2

Executive Summary
The aim of this assignment is to conduct a literature research on the given topic “Machine
Learning in Medicine”. This paper is divided into five sections: first two sections Broad Scan
and Focused Scan shows how the research was carried out to collect the required information
and to make a filter what all is useful.
The third section is completely about the topic. All the information collected regarding the topic
has been presented in this section. Fourth section depicts the layout of the third section and last
section provides an introduction to complete research paper.
After a thorough research on Google scholar, various libraries and IEEE papers two final papers
were chosen to conduct a deep research and to collect the literature review material for the third
section of the document.
1

Table of Contents
1. Broad Scan...................................................................................................................................3
1.1 Research Journal....................................................................................................................3
1.2 Filing System.........................................................................................................................4
1.3 Bibliography for literature selected.......................................................................................6
2. Focused Review...........................................................................................................................6
2.1 Updated Filing System..........................................................................................................6
2.2 Updated Bibliography............................................................................................................6
3. Machine Learning in Medical Diagnosis.....................................................................................7
3.1 Research Problem..................................................................................................................7
3.2 Background Details...............................................................................................................7
3.3 Literature Review..................................................................................................................8
3.4 Summary and Discussion......................................................................................................9
4. Final Outline of Literature Review Chapter..............................................................................10
5. Introduction to Literature Review Chapter................................................................................10
References......................................................................................................................................12
2

1. Broad Scan
I started my research on Google Scholar to collect the most relevant studies regarding the topic
“Machine Learning in Medicine”. I then advanced my research to various libraries such as VU
library and also looked out for various IEEE conference papers online. Based on my research I
found approximately 30 to 40 papers which were in some manner related to the topic of the
study.
1.1 Research Journal
Date Task Action Comment
29/08/2018 A research for
relevant papers for the
topic of study
Entered appropriate
keywords in Google
explorer
More than 10000
papers popped up and
some of them which
looked relevant were
selected
30/08/2018 More research on the
topic
Made a search though
libraries and IEEE
conference papers
Papers which were
related to topic in a
good way were
selected and saved
2/09/2018 Literature Reading Out of the chosen
papers, selected 20 of
them and gave them a
quick reading
Had an understanding
of what machine
learning in medicine
is about
3/09/2018 Literature Reading From the documents
collected on 30th
August, selected
around 30 more
papers and read the
title, looked out for
source, date it was
published on.
Got more relevant
papers for the
assignment
3

4/09/2018 Filtration of selected
Journals
Based on the quick
reading and type of
document , date it was
published on, title of
the document
Got more relevant
papers for the
assignment
7/09/2018 Selection of two most
relevant papers
Selected five most
relevant papers of
focused review and
read them thoroughly
Decided on 2 papers
to continue with
writing assignment
8/09/2018 Filled out the filing
journals for both
broad scan section
and focused review
section
Made a table of all the
papers collected and
their source
No remarks
9/09/2018 Started with the
section 3
Based on first paper
out of 2 selected
papers. Prepared and
written the section 3
All information from
the paper was
paraphrased in my
own words
10/09/2018 More on section 3 Based on second
paper out of 2 selected
papers. Prepared and
written the section 3
All information from
the paper was
paraphrased in my
own words
11/09/2018 Completed section 4
and 5
Based on section 3
made an outline and
typed it
Completed the
Assignment before
deadline
1.2 Filing System
Source Keywords Used Number Returned
Literature
Number collected
Literature
Search Engine Google Machine learning in 78965 10
4

medicine
Machine learning in
Pharma
Medical diagnosis
though Machine
Learning
9877
4563
4
3
Google Scholar Machine learning in
Pharma research
papers pdf
Breast cancer MRI
diagnosis
26745
8056
8
5
IEEE Benefits of Machine
learning in Healthcare
Need For Medical
Specialization In
Machine Learning
345
786
3
4
VU Library Role of machine
learning in medicinal
science
Machine learning in
medicine
56784
2320
2
1
5

1.3 Bibliography for literature selected
Azar, A.T. and Metwally, S.M. (2013). Decision tree classifiers for automated medical diagnosis,
Neural Comput. Appl., vol. 23, no. 7–8, pp. 2387–2403.
Barros, R. C. and Basgalupp, M. P. (2014). Evolutionary Design of Decision-Tree Algorithms
Tailored to Microarray Gene Expression Data Sets, IEEE Trans. Evol. Comput., vol. 18, no. 6,
pp. 873–892.
Breiman, L. (2006). Bagging predictors. Mach Learn;24:123–140.
Chang, C.L. and Chen, H. (2009). Applying decision tree and neural network to increase quality
of dermatologic diagnosis, IEEE Research Paper, vol. 36, no. 2, Part 2, pp. 4035–4041, Mar.
2009.
Çınar, M., Engin, M., Engin, Z. and Ziya, Y. (2009). Early prostate cancer diagnosis by using
artificial neural networks and support vector machines, Expert Syst. Appl., vol. 36, no. 3, Part 2,
pp. 6357–6361, Apr. 2009.
Corren, J. (2011). Lebrikizumab treatment in adults with asthma. N Engl J Med; 365:1088–1098.
Cruz R., Mesa, A., Carrillo, C. and Barrientos, M. (2009). Discovering inter observer variability
in the cyto-diagnosis of breast cancer using decision trees and Bayesian networks, Appl. Soft
Comput., vol. 9, no. 4, pp. 1331–1342.
Das, A. and Bhattacharya, M. (2009). A Study on Prognosis of Brain Tumors Using Fuzzy Logic
and Genetic Algorithm Based Techniques, in International Joint Conference on Bioinformatics,
Systems Biology and Intelligent Computing, IJCBS ‟09, pp. 348–351.
Deepa, S. N. and Devi, B.A. (2011). A Survey on Artificial Intelligence Approaches for Medical
Image Classification, Indian J. Sci. Technol., vol. 4, no. 11, pp. 1583–1595.
Delen, G., Walker, D. and Kadam, A. (2015). Predicting Breast Cancer Survivability: A
Comparison of Three Data Mining Methods, Artif Intell Med, vol. 34, no. 2, pp. 113–127.
Esfandiari, N., Babavalian, M.R., Moghadam, R. and Tabar, V.K. (2014). Machine learning in
medicine, Expert Syst. Appl., vol. 41, no. 9, pp. 4434–4463.
6

Freitas, D. C., Wieser, A. and Apweiler, R. (2010). On the importance of comprehensible
classification models for protein function prediction, IEEEACM Trans. Comput. Biol.
Bioinforma. IEEE ACM, vol. 7, no. 1, pp. 172–182.
Guyon, I. and Elisseeff, A. (2013). An introduction to variable and feature selection. J Mach
Learn Res; 3:1157–1182.
Han, M., Kamber, J. and Pei, J. (2011). Data Mining: Concepts and Techniques, 3rd Revised ed.
Burlington, MA: Morgan Kaufmann Publishers In.
Hassanien, A.E. and Kim, T. (2012). Breast cancer MRI diagnosis approach using support
vector machine and pulse coupled neural networks, J. Appl. Log., vol. 10, no. 4, pp. 277–284.
Hastie, T., Tibshirani, R. and Friedman, J. (2009). The Elements of Statistical Learning,
Springer Science & Business Media, New York, NY
Ishwaran, H. and Kogalur, U.B. (2008). Lauer MS. Random survival forests. Ann Appl Stat;
2:841–860.
Jagger, C., Andersen, K., Breteler, M.M., Copeland, J.R., Helmer, C. and Baldereschi, M.
(2016). Prognosis with dementia in Europe: A collaborative study of population-based cohorts.
Neurologic Diseases in the Elderly Research Group. Neurology. 54(11 Suppl 5):S16–20. PMID:
10854356.
Kannel, W.B., Doyle, J.T., McNamara, P.M., Quickenton, P. and Gordon, T. (2015). Precursors
of sudden coronary death. Factors related to the incidence of sudden death. Circulation, 51:606–
613.
Komlagan, M., Pan, X., Domenger, J.P., Collins, D.L. and Coupe, P. (2014). Anatomically
Constrained Weak Classifier Fusion for Early Detection of Alzheimer’s Disease. In: Wu G,
Zhang D, Zhou L, editors. Machine Learning in Medical Imaging, p. 141–8.
Kourou, K., Exarchos, T.P., Exarchos, K.P., Karamouzis, M.V. and Fotiadis, D.I. (2015).
Machine learning applications in cancer prognosis and prediction. Comput Struct Biotechnol J;
https://doi.org/10.1016/j. csbj.2014.11.005 PMID: 25750696.
7

Lahsasna, A., Ainon, L.R. Zainuddin, L. and Bulgiba, A. (2012). Design of a fuzzy-based
decision support system for coronary heart disease diagnosis, J. Med. Syst., vol. 36, no. 5, pp.
3293–3306.
Liu, F. and Chen, H. (2013). Inter-modality relationship constrained multi-modality multi-task
feature selection for Alzheimer’s Disease and mild cognitive impairment identification.
NeuroImage; 84:466–75. https://doi.org/10.1016/j.neuroimage. PMID: 24045077 51.
Louridas, P. and Ebert, C. (2016). Machine Learning. IEEE Softw. Sep; 33(5):110–5.
Nguyen, C. Wang, Y. and Nguyen, H.N. (2013). Random forest classifier combined with feature
selection for breast cancer diagnosis and prognostic, J. Biomed. Sci. Eng., vol. 06, no. 05, pp.
551–560.
Ota, K., Oishi, N., Ito, K. and Fukuyama, H. (2013). A comparison of three brain atlases for
MCI prediction. J Neurosci Methods. 221:139–50.
https://doi.org/10.1016/j.jneumeth.2013.10.003 PMID: 24140118
Othman, D. and Yau, T. M. (2007). Comparison of Different Classification Techniques Using
WEKA for Breast Cancer, in 3rd Kuala Lumpur International Conference on Biomedical
Engineering, pp. 520–523.
Ozcift, A. (2011). Enhanced Cancer Recognition System Based on Random Forests Feature
Elimination Algorithm, J. Med. Syst., vol. 36, no. 4, pp. 2577–2585.
Padilla, P., López, M., Górriz, J.M., and Ramírez, K. (2012). Disease Neuroimaging Initiative,
“NMF-SVM based CAD tool applied to functional brain images for the diagnosis of Alzheimer‟s
disease,” IEEE Trans. Med. Imaging, vol. 31, no. 2, pp. 207–216.
Schaefer, G. and Nakashima, T. (2010). Data Mining of Gene Expression Data by Fuzzy and
Hybrid Fuzzy Methods, IEEE Trans. Inf. Technol. Biomed., vol. 14, no. 1, pp. 23–29.
Shilaskar, S. and Ghatol, A. (2013). Feature selection for medical diagnosis : Evaluation for
cardiovascular diseases, Expert Syst. Appl., vol. 40, no. 10, pp. 4146–4153.
8

Suh, G.H. and Shah, A. (2011). A review of the epidemiological transition in dementia—cross-
national comparisons of the indices related to Alzheimer’s disease and vascular dementia. Acta
Psychiatrica Scandinavica; 104(1):4–11. PMID: 11437743.
Suk, H.I. and Shen, D. (2013). Deep learning-based feature representation for AD/MCI
classification. Med Image Comput Comput-Assist Interv MICCAI Int Conf Med Image Comput
Comput-Assist Interv. 16(Pt 2):583–90.
Temurtas, H., Yumusak, N. and Temurtas, F. (2015). A comparative study on diabetes disease
diagnosis using neural networks, Expert Syst. Appl., vol. 36, no. 4, pp. 8610–8615.
Vapnik, V.N. (2009). An overview of statistical learning theory. IEEE Trans Neural Netw;
10:988–999
Vatankhah, V., Asadpour, J. and Fazel-Rezai, R. (2013). Perceptual pain classification using
ANFIS adapted RBF kernel support vector machine for therapeutic usage, Appl. Soft Comput.,
vol. 13, no. 5, pp. 2537– 2546.
Woodruff, P.G., Modrek, B., Choy, D.F. and Jia, G. (2009). T-helper type 2-driven inflammation
defines major subphenotypes of asthma. Am J Respir Crit Care Med;180:388–395.
Zhang, S., Wang, G. and Dong, Z. (2013). An MR brain images classifier system via particle
swarm optimization and kernel support vector machine, Sci. World J., vol. 13.
2. Focused Review
2.1 Updated Filing System
Source Keywords Used Number Returned
Literature
Number collected
Literature
Search Engine Google Machine learning in 78965 7
9

medicine
Medical diagnosis
though Machine
Learning
4563
3
Google Scholar Breast cancer MRI
diagnosis
26745 5
IEEE Need For Medical
Specialization In
Machine Learning
345 3
VU Library Role of machine
learning in medicinal
science
56784 2
2.2 Updated Bibliography
Azar, A.T. and Metwally, S.M. (2013). Decision tree classifiers for automated medical diagnosis,
Neural Comput. Appl., vol. 23, no. 7–8, pp. 2387–2403.
Barros, R. C. and Basgalupp, M. P. (2014). Evolutionary Design of Decision-Tree Algorithms
Tailored to Microarray Gene Expression Data Sets, IEEE Trans. Evol. Comput., vol. 18, no. 6,
pp. 873–892.
Breiman, L. (2006). Bagging predictors. Mach Learn;24:123–140.
Chang, C.L. and Chen, H. (2009). Applying decision tree and neural network to increase quality
of dermatologic diagnosis, IEEE Research Paper, vol. 36, no. 2, Part 2, pp. 4035–4041, Mar.
2009.
Das, A. and Bhattacharya, M. (2009). A Study on Prognosis of Brain Tumors Using Fuzzy Logic
and Genetic Algorithm Based Techniques, in International Joint Conference on Bioinformatics,
Systems Biology and Intelligent Computing, IJCBS ‟09, pp. 348–351.
10

Deepa, S. N. and Devi, B.A. (2011). A Survey on Artificial Intelligence Approaches for Medical
Image Classification, Indian J. Sci. Technol., vol. 4, no. 11, pp. 1583–1595.
Freitas, D. C., Wieser, A. and Apweiler, R. (2010). On the importance of comprehensible
classification models for protein function prediction, IEEEACM Trans. Comput. Biol.
Bioinforma. IEEE ACM, vol. 7, no. 1, pp. 172–182.
Guyon, I. and Elisseeff, A. (2013). An introduction to variable and feature selection. J Mach
Learn Res; 3:1157–1182.
Hassanien, A.E. and Kim, T. (2012). Breast cancer MRI diagnosis approach using support
vector machine and pulse coupled neural networks, J. Appl. Log., vol. 10, no. 4, pp. 277–284.
Hastie, T., Tibshirani, R. and Friedman, J. (2009). The Elements of Statistical Learning,
Springer Science & Business Media, New York, NY
Jagger, C., Andersen, K., Breteler, M.M., Copeland, J.R., Helmer, C. and Baldereschi, M.
(2016). Prognosis with dementia in Europe: A collaborative study of population-based cohorts.
Neurologic Diseases in the Elderly Research Group. Neurology. 54(11 Suppl 5):S16–20. PMID:
10854356.
Kourou, K., Exarchos, T.P., Exarchos, K.P., Karamouzis, M.V. and Fotiadis, D.I. (2015).
Machine learning applications in cancer prognosis and prediction. Comput Struct Biotechnol J;
https://doi.org/10.1016/j. csbj.2014.11.005 PMID: 25750696.
Lahsasna, A., Ainon, L.R. Zainuddin, L. and Bulgiba, A. (2012). Design of a fuzzy-based
decision support system for coronary heart disease diagnosis, J. Med. Syst., vol. 36, no. 5, pp.
3293–3306.
Louridas, P. and Ebert, C. (2016). Machine Learning. IEEE Softw. Sep; 33(5):110–5.
Nguyen, C. Wang, Y. and Nguyen, H.N. (2013). Random forest classifier combined with feature
selection for breast cancer diagnosis and prognostic, J. Biomed. Sci. Eng., vol. 06, no. 05, pp.
551–560.
11

Padilla, P., López, M., Górriz, J.M., and Ramírez, K. (2012). Disease Neuroimaging Initiative,
“NMF-SVM based CAD tool applied to functional brain images for the diagnosis of Alzheimer‟s
disease,” IEEE Trans. Med. Imaging, vol. 31, no. 2, pp. 207–216.
Shilaskar, S. and Ghatol, A. (2013). Feature selection for medical diagnosis : Evaluation for
cardiovascular diseases, Expert Syst. Appl., vol. 40, no. 10, pp. 4146–4153.
Temurtas, H., Yumusak, N. and Temurtas, F. (2015). A comparative study on diabetes disease
diagnosis using neural networks, Expert Syst. Appl., vol. 36, no. 4, pp. 8610–8615.
Vatankhah, V., Asadpour, J. and Fazel-Rezai, R. (2013). Perceptual pain classification using
ANFIS adapted RBF kernel support vector machine for therapeutic usage, Appl. Soft Comput.,
vol. 13, no. 5, pp. 2537– 2546.
Woodruff, P.G., Modrek, B., Choy, D.F. and Jia, G. (2009). T-helper type 2-driven inflammation
defines major subphenotypes of asthma. Am J Respir Crit Care Med;180:388–395.
Zhang, S., Wang, G. and Dong, Z. (2013). An MR brain images classifier system via particle
swarm optimization and kernel support vector machine, Sci. World J., vol. 13.
3. Machine Learning in Medical Diagnosis
3.1 Research Problem
The algorithms of machine learning can fundamentally help in taking care of the medicinal
services issues by manufacturing classified frameworks which can help in assistance to diagnose
and anticipate infectious diseases in their early stages. Conversely, the task of extracting
information from the medicinal data can be challenging depending upon how the data has been
organized. It could be high dimensional, unorganized and heterogeneous and might also possess
outliers and noisy data.
To select the best approach in terms of comprehensibility and accuracy it is must to analyze each
of the available machine learning techniques and validate the performance. In this literature
review, the algorithms of machine learning that has been analyzed for the purpose of accurate
12

medical diagnosis are swarm intelligence, evolutionary algorithms, random forest, support vector
machine and decision tree.
Overview
Machine Learning is an art of learning and making predictions from the past data with the help
of artificial intelligence tools. It achieves this with the help of algorithms or methods developed
to make computer intelligent. It is most useful in those cases where the information domain is in
scarcity. The greatest advantages of using machine learning are scalability, customizability,
speed, automation and accuracy.
3.2 Background Details
Various applications in medicine sector works on the principle of classification algorithms. It is a
two level process. The first level is training phase where a classifier is build using some set of
tuples and the next level is classification phase in which classification is performed on the model
and its performance is evaluated based on the set of tuples defined in phase 1.
Decision Tree Algorithm
This algorithm is categorized as classification algorithm. A tree is constructed on the basis of
divide and conquers strategy. Some set of attributes represents instances. The tree is comprised
of leaves and nodes: where nodes are used to represent various attribute values and leaves to bind
all the nodes in a class structure. Outcome is Boolean.
Support Vector Machine
This algorithm is also categorized as classification algorithm and uses statistical learning
methodology. In Support Vector Machine, hyperplane: an optimal boundary is acquired
autonomously on the probabilistic appropriation of preparing vectors in the set.
Random Forests
Again an example of classification algorithm, Random Forests are best in cases of handling large
amount of data with high accuracy. It is a blend of tree indicators where each tree relies upon the
estimations of an arbitrary vector tested autonomously with a similar conveyance for every one
of the trees in the woods.
13

Evolutionary Algorithm
This is a strategy to find optimal solutions in wide and complex set of databases. This algorithm
is motivated by general evolution: a number of determined candidate solutions which are called
chromosomes are developed with the help of operations like mutation and crossover.
Swarm Intelligence
This is a technology for the solution to real-world problems. In the collected database of
population, behaviors of individuals are analyzed collectively who in turn interacts with each
other inside their own environment within a control system decentralized.
3.3 Literature Review
To classify the medical data is a very complex process and then it is needed to be optimized.
Multiple kinds of machine learning algorithms we learned about in the previous section are
tested many researchers and based on the research different algorithms serves different purposes.
Medical Classification
Due to its simplicity of nature and interpretability, decision algorithm is considered to be the best
option for the purpose of medical classification. It acts as a classifier for the diagnosis of brain
tumor, liver cancer, dermatological diseases and breast cancer. Decision algorithm performances
has been compared to many other classification algorithms on the basis of KNN, Bayesian
Network, logistic regression, case based reasoning and ANN. In breast cancer diagnosis it has an
accuracy rate of 99.5 percent. Regression tree and classification model was proposed by a
researcher Luke which provided differences amongst non-malignant liver tissue and HCC. HCC
is believed to be very dangerous because of its diagnosis at later stages. Decision tree algorithm
helped to discover such hidden patterns and building classification model depending upon it.
Identification of Disease and Generalization
Support vector algorithms provide an advanced generalization to the medical classification
technique. For example: If decision tree algorithm helps in detecting a breast cancer then SVM
helps to detect the type of breast cancer. SVM uses Wisconsin breast cancer diagnosis techniques
14

to detect the cancer and achieves high accuracy. It is also used to successfully detect liver,
diabetes and heart diseases which are generally genetic and fuzzy.
Medical Diagnosis
Four medical datasets: lung cancer, breast cancer, leukemia cancer and colon cancer was
analyzed using random forest algorithm for the purpose of selection of optimal features to
diagnose this dataset and to come up with the best strategically solution. Its accuracy was
compared to 15 other trained algorithms and this has shown 99.87 percent accuracy.
Optimization
The optimization of classifier algorithms are achieved with the help of pre-processing tools such
as ACO and PSO which are part of Swarm Intelligence algorithms. It helps to increase the
accuracy of classification algorithms such as decision tree, random forest etc. and makes sure
that resources have been utilized to minimum.
3.4 Summary and Discussion
Effective execution of machine learning strategies for the purpose of diagnosing medical
diseases can support the association of technology in the health environment. Specifically, in
underdeveloped countries and countries with a high population ration where doctor patient ratio
may rise to 1:1800, machine learning methods for detecting diseases at an early stage could be
very beneficial. Innovation can no chance supplant a doctor’s skill and expertise, however it can
deal with generally clear yet time expending analytic errands and specialists can take up
clinically high demand methodology.
4. Final Outline of Literature Review Chapter
1. Research Problem
2. Overview of Machine Learning
3. Background Details
3.1 Decision Tree Algorithm
15

3.2 Support Vector Machine
3.3 Random Forests
3.4 Evolutionary Algorithm
3.5 Swarm Intelligence
4. Literature Review
4.1 Medical Classification
4.2 Identification of Disease and Generalization
4.3 Medical Diagnosis
4.4 Optimization
3.4 Summary and Discussion
5. Introduction to Literature Review Chapter
The aim of this assignment is to conduct a literature research on the given topic “Machine
Learning in Medicine”. Machine Learning is an art of learning and making predictions from the
past data with the help of artificial intelligence tools. It achieves this with the help of algorithms
or methods developed to make computer intelligent. It is most useful in those cases where the
information domain is in scarcity. The greatest advantages of using machine learning are
scalability, customizability, speed, automation and accuracy.
The algorithms of machine learning can fundamentally help in taking care of the medicinal
services. To select the best approach in terms of comprehensibility and accuracy it is must to
analyze each of the available machine learning techniques and validate the performance. In this
literature review, the algorithms of machine learning that has been analyzed for the purpose of
accurate medical diagnosis are swarm intelligence, evolutionary algorithms, random forest,
support vector machine and decision tree.
16

References
Chang, C.L. and Chen, H. (2009). Applying decision tree and neural network to increase quality
of dermatologic diagnosis, IEEE Research Paper, vol. 36, no. 2, Part 2, pp. 4035–4041, Mar.
2009.
Deepa, S. N. and Devi, B.A. (2011). A Survey on Artificial Intelligence Approaches for Medical
Image Classification, Indian J. Sci. Technol., vol. 4, no. 11, pp. 1583–1595.
17