J. Biomedical Science and Engineering, 2010, 3, 509-516
doi:10.4236/jbise.2010.35071 Published Online May 2010 (http://www.SciRP.org/journal/jbise/
JBiSE
).
Published Online May 2010 in SciRes. http://www.scirp.org/journal/jbise
Advanced decision support for complex clinical decisions
Brain Keltch1, Yuan Lin1, Coskun Bayrak2
1Department of Applied Science, University of Arkansas at Little Rock, USA;
2Department of Computer Science, University of Arkansas at Little Rock, USA.
Received 1 March 2010; revised 11 March 2010, accepted 13 March 2010.
ABSTRACT
A Physician’s decision-making skills are directly re-
lated to the patient’s positive outcomes. Therefore, a
wealth of medical knowledge and clinical experience
are key assets for a physician to have. The goal here
is to use historical clinical data and relationships
processed by Artificial Intelligence (AI) techniques to
aid physicians in their decision making process. Pre-
senting this information in a Clinical Decision Sup-
port System (CDSS) is an effective means to consoli-
date decision results. The CDSS provides a large
number of medical support functions to help clini-
cians make the most reasonable diagnosis and choose
the best treatment measures. Initial results have
shown great promise in accurately predicting Fibro-
sis Stage in Hepatitis patients. Utilizing this tool could
mitigate the need for some liver biopsies in the more
than 170 million Hepatitis patients worldwide. The
prototype is extendable to accommodate additional
techniques (for example genetic algorithms and logis-
tics regression) and additional medical domain solu-
tions (for example HIV/AIDS).
Keywords: Fibrosis; Clinical Decision Support; Decision
Tree; Neural Network
1. INTRODUCTION
1.1. CDSS Definition
In hospital information systems (HIS), there are typically
two main systems: Hospital Management Information
Systems (HMIS) and Clinical Information Systems (CIS)
[1]. HMIS support the hospital administration and trans-
action processing services while the CIS is used to sup-
port the clinical staff activities, to collect and dispose of
clinical medical information, and to accumulate rich
clinical knowledge. The CIS also provide clinical advice,
support clinics, assistant clinical decision-making and to
enhance staff efficiency. Clinical decision support sys-
tems (CDSS) are part of the CIS. It is an information
system which uses expert systems and artificial intelli-
gence (AI) technology to support clinical decision. It
makes integrated diagnostic and medical advice bases on
the collected patients’ information, providing reference
for the clinical medical officers.
1.2. Key Functions
Clinical decision support systems vary greatly in their
complexity, function, and application. A Recent study
[2] on health care information management four key
functions of CDSS were outlined as follows:
1) Administrative: Supporting clinical coding and
documentation, authorization of procedures, and re-
ferrals.
2) Managing clinical complexity and details: Keep-
ing patients on research and chemotherapy protocols;
tracking orders, referrals follow-up, and preventive
care.
3) Cost control: Monitoring medication orders; avoi-
ding duplicate or unnecessary tests.
4) Decision support: Supporting clinical diagnosis
and treatment plan processes; and promoting use of
best practices, condition-specific guidelines, and popu-
lation-based management.
Our project will focus on item four, the decision
support function and, in particular, utilization of his-
torical laboratory data and outcome data processed
through artificial intelligence tools. The combination
of historical data and predictive tools provides valu-
able information in the hands of physicians as they
develop a course of treatment for a patient.
2. BACKGROUND: AI TECHNIQUE IN
CDSS
Decisions about medical treatment are best made by a
trained and experienced physician. These decision mak-
ers can benefit from historical data and artificial intelli-
gence tools. Computer scientists have dreamed of creat-
ing an “electronic brain” [3]. Computer scientists and
doctors alike have been captivated by the potential such
a technology might have in medicine [4]. With intelli-
gent computers able to store and process vast stores of
knowledge, the hope was that they would become per-
fect ‘doctors in a box’, assisting or surpassing clinicians

510 B. Keltch et al. / J. Biomedical Science and Engineering 3 (2010) 509-516
with tasks like diagnosis [3].
Today the importance of diagnosis as a task requiring
computer support in routine clinical situations receives
much less emphasis. The strict focus on the medical set-
ting has now broadened across the healthcare spectrum,
and instead of AI Medical systems, it is more typical to
describe them as CDSS [3].
In our project, we evaluated and compared two tech-
niques that will be core forecasting tools in a CDSS,
which are Data Mining and Neural Networks.
The main purpose of doing data mining and knowledge
discovery on the medical database is to predict disease
and disease classification. Classification and prediction
are two forms of data analysis which can be used to de-
scribe the model of the important data type or predict the
future trends of the data [5].
Commonly used data mining algorithms are: associa-
tion rules, decision trees, rough sets, statistical analysis,
neural networks, support vector machines, fuzzy clus-
tering, Case-Based Reasoning (CBR), Bayesian fore-
casting and visualization technology [6]. The common
methods used in auxiliary diagnosis of clinical disease
are 1) Bayes discriminate analysis 2) artificial neural
network 3) decision tree.
However within the context of the study, the focus
will be concentrated on decision tree and neural net-
work.
2.1. Decision Tree
The decision tree is a very efficient machine learning
classification algorithm. It is the origin in the concept
of learning systems CLS, and then progress to ID3
method. In the end, it evolved to c5.0 which can han-
dle continuous attributes. Well-known decision tree
methods are CART and Assistant [5].
Decision tree learning uses a decision tree as a pre-
dictive model which maps observations about an item
to conclusions about the item’s target value. In these
tree structures, leaves represent classifications and bra-
nches represent conjunctions of features that lead to
those classifications. One of the biggest advantages of
the method is that the learning process does not require
the user to understand a lot of background knowledge
[6].
Nonetheless, data mining is a complex process to
identify the useful information from large data sets.
Although it is common to focus on the development,
analysis and application of algorithms, the data selec-
tion and data pre-processing are the most timecon-
suming activity in the entire data mining process,
which affects the process and results [6].
2.2. Neural Network
Artificial Neural Networks have been proven to build
efficient rule extraction/classification and forecasting
applications. They provide a powerful non-linear ma-
chine learning techniques and are able to extract rele-
vant features from large data sets. They are able to uti-
lize and compare equally quantitative and qualitative
data which is common in the clinical environments [7].
Neural Networks can handle redundant features as wei-
ghts are learned from the training data.
The primary disadvantage of neural networks is that
they will always arrive at a solution for any data set.
This means that the quality of the resulting model is
highly dependent on the quality and breath of the trai-
ning data. It is easy to over-train the model, so that it
can only predict the training data set. This can be mi-
nimized by assuring that convergence stopping stra-
tegies are effective and that the training data set is rep-
resentative of the solutions space. Also, the resulting
neural network solution is simply a node structure with
inter-node weights. This requires validation of the out-
put by additional statistical methods (i.e. decision trees)
that are more understandable by subject matter experts.
3. AI ASSISTED CLINICAL DECISION
SUPPORT SYSTEM
This study will focus on the demonstration and incorpo-
ration of neural network and decision tree techniques
into a Clinical Decision Support System. These two AI
techniques were selected because of their complemen-
tary attributes. Neural networks provide little definition
of their predictive result, while decision tree output pro-
vides clear connections to historical data. Both methods
will provide different information to the physician. Other
AI methods should be considered in future work.
The database we plan on using for this study was col-
lected at Chiba University hospital in Japan, and is a
Practice of Knowledge Discovery in Databases (PKDD)
2005 Discovery Challenge dataset [11]. The data set
contains patient data, laboratory data, and liver biopsies
data on 771 hepatitis B and C patients. The goal will be
to evaluate whether laboratory examinations can be used
to estimate the stage of liver fibrosis. If this is possible,
physicians may be able to use laboratory examinations
as substitutes for biopsies and to aid in the treatment
scenario. Liver biopsy is an invasive procedure and en-
tails risk to patients. Decision tree and neural network
methods based on a historical dataset will aid physicians
in the development of treatment plans for Hepatitis pa-
tients.
The overall architectural representation of the system
is shown in Figure 1.
The system is designed to be utilized by a physician to
assist them in the development of a treatment plan for
Hepatitis B and Hepatitis C patients. This system ad-
dresses an important question for the treatment – “Sho-
uld a liver biopsy procedure be conducted?” This proce-
dure provides important information on the advancement

B. Keltch et al. / J. Biomedical Science and Engineering 3 (2010) 509-516 511
Raw
Laboratory
Data
Patient
Demographic
Data
Data
Cleansing
& Processing
Knowledge
Base
Inference
Engine
Decision
Tree
Neural
Network
USER
INTE RFACE
Patient
Data in
Risk, Treatmen t Suggestion s
For Patient
Prototype Scope
Remark
Figure 1. AI assisted clinical decision support system.
of the disease and fibrosis formation which are the key
outcome measures. These fibrosis stage results deter-
mine the treatment protocol. This study attempts to pre-
dict fibrosis stage base on laboratory and patient data.
The methodology and approach of processing data, ap-
plying AI techniques, and development of the resulting
knowledge base can be utilized as a pattern for other
medical treatment needs represented in a CDSS.
3.1. Data Processing and Cleaning
The sample hepatitis data set for our study is derived
from the ECML/PKDD 2005 Discovery Challenge
found at [11]:
The hepatitis dataset contains the results of laboratory
examinations taken on the patients of hepatitis B and C,
who were admitted to Chiba University Hospital in Ja-
pan. Hepatitis A, B and C are virus infections that affect
the liver of the patient. Hepatitis B and C chronically
inflame the hepatocyte, whereas hepatitis A acutely in-
flames it. Hepatitis B and C are especially important
because they have a potential risk of developing liver
cirrhosis or hepatocarcinoma. An indicator that can be
used to know the risk of cirrhosis or hepatocarcinoma is
fibrosis of hepatocyte. For instance, liver cirrhosis is
characterized as the terminal stage of liver fibrosis.
We utilized three tables for our study, as show in Ta-
bles 1-3, below:
The MID field provided a common link between the
three tables. Since the question we were addressing is to
evaluate whether laboratory examinations can be used to
estimate the stage of liver fibrosis our goal was to obtain
one table that contained patient information, fibrosis
stage and laboratory data. We chose to utilize in-hospital
laboratory data, as it was more complete and we felt it
would have better controls on quality and consistency.
Because of the large volume of data all files were
translated into MS Access tables to process. The follow-
ing steps were utilized:
1) Select all records from Table 2 - 649 records.
2) Link to Table 1 based on Patient ID - 649 records.
3) Link to the Laboratory Examination Table 3 based
on Patient ID. Select records where laboratory examina-
tions were available for at least 90% of our biopsy pa-
tients, resulting in 16 examination parameters.
4) Link to and Match appropriate lab examination re-
sults (Table 3) with patient information and fibrosis
stage information (the combined Tables 1 an d 2). Select
the laboratory examination that was earlier that the biopsy,
but not more than 30 days earlier. This resulted in 425 re-
cords with complete data on the 16 laboratory examination
results, patient information, and fibrosis stage biopsy re-
sults.
Table 1. Basic information of patients: pt_e030704.csv (total
771 records).
Item Meaning
MID Identification of the
patient (masked) unsigned integer
PT Sex Sex of the patient M(male) or F (female)
PT BirthDateBirth date of the patient YYYYMMDD
Table 2. Results of biopsy: bio_e030704.csv (total 694 re-
cords).
Item Meaning Remark
MID Identification of the
patient (masked) unsigned integer
BIOPSY Lo_No Identification of the
specimen unsigned integer
BIOPSY Hepati-
tis_Type Type of hepatitis B or C
BIOPSY Ex-
am_Date
The date when biopsy
was performed YYYYMMDD
BIOPSY Hepati-
tis_Subtype Subtype of Hepatitis
(Data concerning
this field were
dropped for sim-
plicity)
BIOPSY Facility
Name of the facility
where the specimen
was collected
(Data concerning
this field were
dropped for sim-
plicity)
BIOPSY FibrosisBiopsy report about
progress of fibrosis
Discrete values:
0(F0; no fibro-
sis)-4(F4; severe)
BIOPSY ActivityBiopsy report about
activity of virus
Discrete values:
0(A0; no activ-
ity)-3(A3; severe),
or FALSE (activ-
ity could not be
specified)
Note that FALSE
does NOT repre-
sent ‘no activity’,
but represent
‘activity could not
be specified’.
According to the
donator, FALSE
should be treated
as a missing value.

512 B. Keltch et al. / J. Biomedical Science and Engineering 3 (2010) 509-516
Table 3. Results of in-hospital examinations: ilab_e030704.csv.
Item Meaning Remark
MID Identification of the
patient (masked)
ILAB Exam_Date Date of the examination YYYYMMDD
ILAB Exam_No
ID for each examination
performed repeatedly on
the same day (n-th ex-
amination)
ILAB Exam_Name Name/code of the ex-
amination
ILAB Ex-
am_Result
Result of the examina-
tion
3.2. Methods and Analysis
The objective is to use the decision tree and neural net-
works to predict fibrosis stage from patient data and la-
boratory data. We have a data set (see data preparation)
of 424 historical that we will use to train and validate a
decision tree model and a neural network model. Table 4
is a map of the data we will be using.
3.2.1. Decision Tree Analysis
Data Mining is an analytic process designed to explore
data (usually large amounts of data - typically business
or market related) in search of consistent patterns and/or
systematic relationships between variables, and then to
validate the findings by applying the detected patterns to
new subsets of data. Decision Tree’s are one methodol-
ogy utilized in the data mining set of tools. The ultimate
Table 4. Data utilized for neural network.
INPUT LABORATORY DATA
Abbreviation Description
ALB Albumin
ALP alkaline phosphatase
CHE Cholinesterase
CL Chloride
CRE Creatinine
D-BIL bilirubin, direct
G-GTP gamma-glutamyltranspeptidase
G.GL gamma-globulin
I-BIL bilirubin, indirect
K Potassium
T-BIL bilirubin, total
T-CHO cholesterol, total
TP protein, total
TTT thymol turbidity test
UA uric acid
UN blood urea nitrogen
INPUT PATIENT DATA
Abbreviation Description
AGE Patient Age at Time of Test
SEX Male = 1 or Female = 2
HEP_B_C Hepatitis B = 1 or Hepatitis C = 2
OUTPUT FIBROSIS STAGE DATA
Abbreviation Description
FIBSTAGE Fibrosis stage (0 to 4)
goal of data mining is prediction - and predictive data
mining is the most common type of data mining.
For the project, Weka 3.4 [9] which is a collection of
machine learning algorithms for data mining tasks is
used. The algorithms can either be applied directly to a
dataset or called from your own code. Weka contains
tools for data pre-processing, classification, regression,
clustering, association rules, and visualization. It is also
well-suited for developing new machine learning schemes.
It also provides Decision Tree capability that we will use
in this study.
3.2.1.1. Decision Tree Testing Procedure
Step 1: Data preprocessing
The input data file needs to be organized in the form
of ARFF in order to be processed in the Weka environ-
ment. In the file, all the values for the attributes needed
to be filled in. If there is any missing value for an attrib-
ute, a “?” is used for substitution.
The Figure 2 shows the attributes and the patient’s
information whose Masked ID is 1. In the input file,
even the data for the predicted attribute Biopsy Fibrosis
needs to be completed for the training data set. For the
analysis 392 patients’ information was used to develop
the decision tree module and predict the other 32 pa-
tients’ fibrosis stage.
The remaining 32 patients’ data was formatted into a
test ARFF file. In this file the fibrosis is substituted with
a “?”. Figure 3, below shows the patient’s information
whose MID is 907.
Step 2: Build the Model and get Decision Tree
We used the C4.5 algorithm to construct the decision
tree. The root node is the Biopsy Fibrosis with five
branches which present the 5 levels where the different
laboratory values and patient characteristic are assigned
one level at a time. Figure 4 shows a graphical presenta-
tion of the complete decision tree model of the data set.
The Figure 5 shows a graphical presentation of the
branch of the decision tree model where CHE <= 4.48.
Figure 2. The WEKA training data file.

B. Keltch et al. / J. Biomedical Science and Engineering 3 (2010) 509-516 513
Figure 3. WEKA test data file.
Step 3: Prediction
The result of predicting the values with the con-
structed decision tree model is shown in Table 5. The
table contains 32 patients’ fibrosis stage values; the first
column is the actual fibrosis stage, α, from the biopsy;
the second is the predict values, β, by the decision tree
model; and the third column shows the difference, γ = α – β.
The decision tree model results showed an accuracy of
37.5% (12/32) of correct fibroid prediction. Predicting
fibrosis values within a range of +or – one of the actual
fibroid stage showed an accuracy of 91% (29/32).
3.2.1.2. Suggestions to Improve Decision Tree
Accuracy
1) An increase in the number of training data exam-
ples will increase the correctness. Hence we can get a
more powerful model, if we have a comprehensive
training data set.
2) It is possible that obtaining additional laboratory
attributes will increase the correctness. We were able to
use only 16 laboratory attributes to predict the result.
This was limited by the original data availability. We
deleted some laboratory data since the incomplete data
may also affect the test result.
3) The parameter setting for the decision tree algo-
rithm utilized by Weka software is entered manually. We
used the default value for the test because of limited time
for the analysis. We may be able to improve results by
testing additional parameters.
3.2.2. Neural Network Analysis
Artificial neural networks (ANNs) are systems that are
constructed to use some organizational principles re-
sembling those of the human brain. They are information
processing systems that demonstrate the ability to learn,
recall, and generalize from training patterns or data.
ANNs are good at tasks such as pattern matching and
classification, data clustering, and forecasting.
Table 5. Decision tree prediction quality.
Known Fibrosis
Stage (α)
Predicted Fibrosis
Stage (β)
Difference = (Known-
Predicted) γ = α – β
10 1
10 1
22 0
22 0
1 1 0
43 1
32 1
32 1
12 -1
11 0
10 1
01 -1
3 3 0
24 -2
11 0
21 1
11 0
11 0
43 1
23 -1
1 2 -1
21 1
23 -1
11 0
11 0
14 -3
21 1
22 0
1 3 -2
11 0
10 1
12 -1
For the experiment we used a freeware tool called
Neuro 3 [10] which uses the back propagation neural
network (BPN). The term backpropagation refers to the
training method by which the weights of the network
connection are adjusted. The calculations procedure is
feedforward, from input layer through hidden layers to
output layer. During training, the calculated outputs are
compared with the desired values, and then the errors are
backpropagated to correct all weight factors.
All Training factors are defined by the users, including
 Number Hidden Layers
 The Threshold Value
 Transfer functions
 Learning Rate
 Momentum Coefficient
 Maximum Iterations
 Convergence Criteria
Network and training parameters are stored in sce-
nario files and on projects spreadsheets which may be
copied and pasted to other spreadsheet programs.
3.2.2.1. Neural Network Testing Procedures
The process for testing the predictive capabilities of the
Neural Network includes the following steps:
Step 1. Divide the historical data set into a Training
Data Set and a Prediction Data Set.

514 B. Keltch et al. / J. Biomedical Science and Engineering 3 (2010) 509-516
Figure 4. Decision tree.
Figure 5. Section of the decision tree.
JBiSE

B. Keltch et al. / J. Biomedical Science and Engineering 3 (2010) 509-516 515
Step 2. Define the Neural Net architecture. The num-
ber of input nodes is defined by our 19 input (three pa-
tient data parameters and 16 laboratory parameters) pa-
rameters. And we have one output node (fibroid stage).
The number of nodes in the hidden Layer is shown in the
Figure 6.
Step 3. Define the run parameters. For this study we
will only adjust the number of iterations that the error
will be calculated and the weight adjusted. This error
calculation and backpropagation will be executed 10,000
or 30,000 times.
Step 4. Run the neural network model on the training
data set. The Neuro 3 tool provides a R-squared value
and Sum of Errors for the training data set that provides
some indication of the goodness of the model. A exam-
ple of the Neuro 3 screen is shown in Figure 7.
Step 5. Utilize the trained neural network in Neuro 3
to predict the fibrosis stage for the prediction data set.
Evaluate the goodness of fit by evaluating the percent of
predictions that were correct, and that were within plus
or minus 1, 2, or 3.
Step 6. Adjust the parameters in Step 1 through Step 5
until the optimal fit is achieved.
Step 7. Utilize the final selected model and train this
model with all 424 historical data elements. Identify any
deficiencies in the process or items that would be helpful
for future work.
3.2.2.2. Testing and Validation Results
Eleven neural network models were generated. The best
Figure 6. Architecture of a BPN neural network.
Figure 7. Neuro 3 screen.
fitting model is Run Number 8, a 19-4-1 architecture
trained with 10,000 iterations. This model provides fair-
ly good predictive capabilities with 56% of predictions
being correct (γ= α – β = 0) and 90% being within +/-
one fibroid stage (γ= α – β = ± 1). The run statistics are
show below in Table 6.
The process of testing and validation of results re-
vealed the following future considerations:
First, the increase in the amount of training data
would greatly improve the predictive capability of the
tool. This can be observed comparing Run Number 1 to
Run Number 6. The application should allow for inclu-
sion of additional data.
Second, additional data elements may increase the
predictive capabilities of the tool. We only had sufficient
data coverage for 16 laboratory data elements. Addi-
tional laboratory data elements should be acquired and
evaluated utilizing these techniques.
Third, the analysis tools and selection of neural net-
work architecture and run parameters was selected ma-
nually. This was a limitation of time available for this
study. Automated analysis techniques and hybrid tech-
niques combining neural network and genetic algorithms
should be considered. An illustration of the sensitivity of
predictive capabilities to hidden layer nodes shows in the
Figure 8 below. The vertical axis shows the percent of
patients that the neural network predicted the correct
fibrosis stage. The horizontal axis shows the number of
hidden nodes specified in the neural network model,
indicating that the accuracy of prediction is very sensi-
tive to the number of hidden nodes in the model.
4. CONCLUSIONS
We have completed the development of a prototype that
utilizes publically available patient data to address a sin-
gle clinical decision–the prediction of fibrosis stage
shown in Figure 9. The near-term customers for our
project prototype are clinicians treating Hepatitis B and
C patients.
Currently the key diagnostic tool in assessing the de-
gree of liver disease in these patients is a liver biopsy.
This procedure is invasive and requires the physician to
Figure 8. Correctness as a function of the number of hidden
nodes.

516 B. Keltch et al. / J. Biomedical Science and Engineering 3 (2010) 509-516
extract a small amount of the patient’s liver using a fine
needle and a suction syringe. While complications rates
are low, < 3%, when they do occur they have large im-
pacts. Also these procedures are expensive ($ 2,000 - $
4,000 per test). Within the past several years several
non-invasive tests have been developed as alternatives to
biopsies. These tests generally require use of ultra sound
equipment and/or specialized/proprietary blood test.
These include: FibroScan, FIBROSpect II, FibroTest,
FibroTest-ActiTest, and HCV-FibroSure. Cost for these
test range from $ 400 to $ 700 [8].
The advantage of our method is that only standard
liver panel blood tests are required to make the predic-
tion of liver fibrosis. This provides low cost testing
without any additional impact to the patient. Beyond the
specific application in our prototype, linking AI tools
Figure 9. Prototype web application.
Table 6. Run statistics, run 8 is used for prediction.
Run
Number
Neural
Network
Architecure
Training
Data Set
Number
Prediction
Data Set
Number
Numer of
Iterations
Training
R-squared
Training
Sum of
Errors
Prediction
% Correct
Prediction
% +/- 1 Prediction % +/- 2 Prediction +/- 3
1 19-8-1 300 124 10,000 0.81 25.5 30 71 92 100
2 19-8-1 300 124 30,000 0.86 19.6 27 72 84 97
3 19-6-1 300 124 10,000 0.72 32.2 38 76 90 99
4 19-6-1 374 50 30,000 0.66 43.1 40 76 92 98
5 19-6-1 374 50 10,000 0.68 39.3 34 74 92 98
6 19-8-1 374 50 10,000 0.74 34.7 34 82 98 100
7 19-10-1 374 50 10,000 0.78 31.3 28 64 94 98
8 19-4-1 374 50 10,000 0.56 50.3 56 90 96 100
9 19-2-1 374 50 10,000 0.39 64.1 34 90 100 100
10 19-5-1 374 50 10,000 0.62 45.1 48 84 96 100
11 19-3-1 374 50 10,000 0.47 57.4 44 88 98 100
with clinical decision support systems has wide applica-
bility for other medical and healthcare solutions.
[3] Enrico, C. (2003) Guide to health informatics. 2nd Edition,
Chapter 19, Artificial Intelligence in Medicine: An In-
troduction. Hodder Arnold, Arnold.
We have utilized Visual Studio 2008 for our prototype.
With continued modifications beyond the prototype, we
will utilize SQL Server 2008. Additionally we are utiliz-
ing two open source analytic tools for the AI component
of our project. Namely, we are using “Neuro 3” a Visual
Basic application for the neural network application and
“Weka” for the decision tree portion of our project.
[4] Robert, S.L. and Lee, B.L. (1959) Reasoning foundations
of medical diagnosis: Symbolic logic, probability, and
value theory aid our understanding of how physicians
reason. American Association for the Advancement of
Science, 130(3366), 9-21.
[5] Špečkauskienė, V. and Lukoševičius, A. (2009) Method-
ology of adaptation of data mining methods for medical
decision support: Case study. Data Mining, 9(2), 228-
235.
REFERENCES [6] Han, J. and Kamber, M. (2001) Data mining concepts
and techniques. 2nd Edition, Academic, Kluwer.
[1] Begg, R. (2009) Artificial intelligence techniques in
medicine and health care. Concepts, Methodologies, Tools,
and Applications, Sugumaran, V., Ed., 48(12), 1-99. [7] Taylor, J.G. (1996) Neural networks and their applications.
John Wiley and Sons, New York.
[2] Perreault, L.E. and Metzger, J.B. (1999) A pragmatic
framework for understanding clinical decision support. In:
Middleton, B., Ed., Clinical decision support systems.
Journal of Healthcare Information Management, 13(2),
5-21.
[8] http://janis7hepc.com/biopsies1.htm
[9] http://www.cs.waikato.ac.nz/ml/weka/
[10] http://www.keltch.com/neuro3.html
[11] http://lisp.vse.cz/challenge/ecmlpkdd2005/
JBiSE