Dynamic Context Adaptation for Diagnosing the Heart Disease in Healthcare Environment Using Optimized Rough Set Approach

International Journal on Soft Computing (IJSC) Vol.3, No.2, May 2012
DOI: 10.5121/ijsc.2012.3202 23
DYNAMIC CONTEXT ADAPTATION FOR
DIAGNOSING THE HEART DISEASE IN
HEALTHCARE ENVIRONMENT USING OPTIMIZED
ROUGH SET APPROACH
G.NaliniPriya1
, A.Kannan2
and P.Ananahakumar3
1
Research scholar ,Department of IST, Anna University, Chennai
nalini.anbu@gmail.com
2
Professor, Department of IST, Anna University, Chennai
3
AssociateProfessor, Department of IT, Anna University, Chennai
ABSTRACT
In a smart medical environment, the ability to adapt context is critical. Context adaptation (CA) requires
applications to delegate adaptation control to an entity that can receive state information and trigger
adaptation in multiple applications. Thus, the idea of CA is to extract general knowledge from each
contextualized system or medical dataset and to use it, in adapted form, when facing a similar problem in a
different environment. In this paper, we describe a model for providing smart intelligent diagnosing
services to the healthcare Environment and present work on our own architecture that has been
designed to meet the key requirements of healthcare in a context-aware adaptive applications. The
proposed system is capable of providing context intelligence, Decision Making, adaptability and extensible
application framework for dynamic remote healthcare environments. The objective of the research is to
revolutionize daily human life by making people’s surroundings flexible and adaptive.
KEY WORDS
Context adaptations, CAD(coronary artery disease),Heart disease dataset , Context Aware Services,
context service discovery, rough set, Fuzzy set.
1. INTRODUCTION
The context [4] is defined as the interrelated conditions in which something exists or occurs. This
means that context is the mutual relationship between the many conditions that exist in a given
situation in which some actor exists or an event occurs. Context-aware [19] applications look at
the who’s, where’s, when’s and what’s (that is, what the activities are occurring) of entities and
use this information to determine why a situation is occurring. An application does not actually
determine why a situation is occurring, but the designer of the application does .The designer uses

24
incoming context to determine why a situation is occurring and uses this to encode some action in
the application.
1.1 General structure of context
The structure of the context mainly has User, Social environment, Task, Condition, Infrastructure,
Location categories. When compare together these two lists of categories resembles each other
very much. But there are minor differences such as time which is not apparent in the list given by
[6]Schmidt et al. and the more differentiated physical environment which is in the list by Dey[5]
et al. is captured in one item e.g. Location.All actions carried out by a human are taking place in
context - in a certain situation. All contexts and situations are embedded in the world, but the
perception of the world is dictated by the immediate context someone is in. Explicit user
interaction with an application is embedded into the context of the user and is also a way of
extending the context of the user, e.g. by having access to the [6]network.
Figure 1.Implicit Human Computer Interaction Model
1.2 About the Heart disease Dataset
Heart disease, which is usually called coronary artery disease (CAD), is a term that can
refer to any condition that affects the heart. CAD is a chronic disease [14][15] in which
the coronary arteries gradually hardens and narrow. It is the most common form of
cardiovascular disease. Moreover, cardiovascular disease is the biggest killer of the world.
While many people with heart disease have symptoms such as chest pain , fatigue and
breathing discomfort etc.. The data generally used for diagnosing the CAD will be
multivalent in nature. Having so many factors to analyse to diagnose the heart diseases,
physicians generally make decisions by evaluating the current test results of the patients.
The previous decisions made on other patients with the same condition are also examined
by the physicians. These complex procedures are not easy when considering the number
of factors that the physician has to evaluate. So, diagnosing [16] [17] the heart disease of
a patient involves experience and highly skilled physicians. In this paper we have taken
few of the major symptoms of Heart disease and trying to give solution through Context
awareness and Rough set with our own Architecture.

25
1.3 Context Adaptation
Context adaptation (CA) is a knowledge engineering technique that deals with the instantiation of
a universal model to a context-dependent system [1][2]. While the universal model usually
represents versatile knowledge that can be reused in several different environments, the context-
dependent system is a specialized version of the model that is properly suited to work in a specific
environment. The use of CA is motivated by the observation that many real-world problems are
common to different environments, exhibiting independent and similar patterns. Thus, the idea of
CA is to extract general knowledge from each contextualized system and to use it, in adapted
form, when facing a similar problem in a different environment.
1.4 Basic Rough Set Theory
The main concept of rough set [13] theory is an indiscernibility relation associated with a set of
attributes. Cases with the same attribute profile are grouped to form elementary sets. Cases within
each elementary set are indiscernible. Any finite union of elementary sets is a definable set. If a
set of attributes does not contain any redundant information (no redundant attributes), it is a
minimal set or a reduct. Once reducts are established, rules are generated linking the variables to
the outcome on a development sample. These rules are then applied to new data and the
classification rate is determined and the relative importance of variables to outcomes can be
discerned.
Rough Set Theory deals with information represented by a table called an information system,
which consists of objects and attributes. An information system is composed of a 4-tuple as
following S = (U, A, V; f) where U is the universe, a finite set of N objects where {X1, X2, .., N},
A = C U D is condition attribute V is attribute value. f: U x A - V is the total decision function
called the information function. upper and lower approximation is an important concept in Rough
Set Theory.
2. PROPOSED WORK
In this paper, we define a new context Architecture model named DYCOD (Dynamic context
Adaptation).This Model will provide a dynamic context adaptation for CAD (coronary artery
disease) Dataset using Roughest Theory. We divide the entire execution into problem of context
adaptation and problem solution through roughset Theory. Here context awareness is combined
with the service discovery mechanism to render better adaptive context aware services to the
CAD (coronary artery disease) Patients.
2.1 The Dycoad Generic Model Approach
The Proposed work starts with user needs. The context information (value) is detected from the
context communication devices. This will be recorded in the administration department and the
context values are classified according to the priority of the request. Then the attributes may be
checked with validity through reasoning algorithm. The history of the context data is analyzed
and stored in the context repository. The information will be retrieved based on the request. The
Value of the attribute check for completeness i.e. all the value will be ensuring that it contains full

26
set of value for the each set of contexts. If any of the value is missing in the attribute, it is treated
as a partial context. Then it’ll be solved with Roughest[5][7] Algorithm. If the value is perfectly
match with context data previously stored, then the requested context may be provided to the
user.
The objective of the proposed work is to revolutionize the human life by making people flexible
and adaptive .In this paper the roughset theory and contexts adaptation is experimented in the
guise of healthcare application especially for Heart disease patients. The key goal of the Dycoad
architecture is to sense the context, understand the context and retrieve services, which might be
useful or relevant to the patient needs, and which are ranked by the relevance for the individual
patient’s special need.
Figure. 2 Simple layout of Dycoad Architecture.
The requested services are provided according to the user need with respect to contexts(eg
person context, disease context, place context….) The context aware service [29][30]model
Dycoad find the right services when they are needed, either on demand or when new services are
provided. This is like an integrated automated information push and pulls cycle.
Context Discovery
Reasoning
Algorithm
Context
History
Context
Repository
CAS
Management
Polling, Adv
Selection
ConDec
Context Distributor
RST Algorithm
N
O
Partial
Context
YES
USER
NEED
CAS-Context Aware Service Management, ConDec-Context Decision Provider
Check
services
USE
CONTEXT

27
A successful Dycoad context model of health information is proposed here to achieve the good
contextual knowledge over the health domain. The recent convergence of ubiquitous computing
and context-aware computing has seen a considerable rise in interest in various context-aware
applications especially in medical environment based on the context. Context information
gathered from various Context[8][9][10] aware Service directory is the basis for enmeshing
ubiquitous computing into our daily lives and exhibiting the autonomy of applications. Storing
and acquiring such information in a single smart controller can be easily handled by a centralized
context repository. However, handling large-scale context information over multiple smart
controllers requires appropriate context lookup architecture. Distributing objects to multiple
controllers in a wide-area network does provide a scalable and reliable solution.
2.2 Rough set Theory Based Approach for Context Adaptation
Inconsistent data sets are handled by rough set theory using lower and upper approximations for
every concept. These approximations are definable using existing attributes. Furthermore, from
the concept lower and upper approximations certain and possible rule sets are induced using one
technique of data mining called rule induction. Generally, input data sets are frequently
incomplete, i.e., some attribute values are missing. In other words, corresponding decision tables
are incomplete.
In general, in data mining two approaches are used to handle incomplete data. The first one is pre-
processing of input data and then main processing of data mining such as rule induction.
Typically, pre processing deals with replacing the missing attribute values by the most common
value, ignoring cases with missing attribute values, etc. The second one is _knowledge is acquired
directly from incomplete data sets taking into account that some attribute values are missing.
Typical systems using this approach are C4.5 and MLEM2.
In this paper rule induction approach is performed directly from incomplete data. Furthermore, it
is assumed that there are two reasons for data to be incomplete. The first reason is that an attribute
value, for a specific case, is lost. This may happen when the original value was erased or
mistakenly not included into the data set. The second reason for incompleteness is based on the
lack of relevance. For example, it was possible to diagnose a patient using only selected results of
tests (attributes), while other test results were redundant. Such missing attribute values will be
called do not care conditions.
2.3 Decision Making Using Rough Set
It is assumed that in the same decision table some attribute values are lost and some are do not
care conditions [15] from the view point of rough set theory where a method for rule induction
was introduced in which each missing attribute value was replaced by all values from the domain
of the attribute.
Originally such values were replaced by all values from the entire domain [11][12] of the
attribute, selected results of tests (attributes), while other test results were redundant. Such
missing attribute values will be called do not care conditions.

28
Figure. 3 Decision making using rough set
For incomplete decision tables, for a given characteristic relation and the concept, there are three
different possible ways to define lower and upper approximations, called singleton, subset, and
concept approximations. As it was observed in singleton lower and upper approximations are not
applicable in these Dycoad Architecture blocks of attribute-Value Pairs, Characteristic Sets, and
Characteristic Relation: In the sequel it is assumed that all decision values are specified, i.e., they
are not missing. Also, it is assumed that all missing attribute values are denoted by ? and by*,
lost values will be denoted by ? and do not care conditions will be denoted by *.
Additionally, another assumption for each case is at least one attribute value should be specified.
2.4 Case study – Decision making from incomplete data
For incomplete decision tables [12]13][14]1the definition of a block of an attribute-value pair
must be modified. For the example of an incomplete data set from Table 1.
· If an attribute a there exists a case x such that (x, a) = ? i.e., the corresponding value is lost, then
the case x should not be included in any block [(a, v)] for all values v of attribute a.
·If for an attribute a there exists a case x such that the corresponding value is a do not care
condition, i.e., (x, a) = *, then the corresponding case x should be included in blocks [(a, v)] for
all specified values v of attribute a.
[(Chest Pain, high)] = {1, 4, 5, 8},
[(Chest pain, very high)] = {2, 8},
[(Chest pain, normal)]= {6, 7, 8},
Notification
Processing
Data
Sensor
Data
RFID
PDA
/Mobile
Laptop
Rough set
(variable sets and
rules)
Predominant
Situations
Individuali-zation Context
Analysis
Rule Repository
Notification
Context
Applications
Context
Repository
Context Management Reasoning-Decision component

29
[(Breathing Discomfort, yes)] = {2, 4, 6, 8},
The characteristic set KB(x) is the intersection of blocks of attribute-value pairs (a, v) for all
attributes a from B for which_(x, a) is specified and (x,a) = v.
Table 1.Incomplete decision table
KA(1) = {1, 4, 5, 8} ∩ {1, 3, 6, 8} = {1, 8},
KA(2) = {2, 8} ∩ {2, 4, 6, 8} ∩ {2, 4, 5, 7, 8} = {2, 8},
KA(3) = {3, 7} ∩ {1, 3, 6, 8} = {3},
KA(4) = {1, 4, 5, 8} ∩ {2, 4, 6, 8}∩{2, 4, 5, 7, 8} = {4, 8},
KA(5) = {1, 4, 5, 8} ∩ {2, 4, 5, 7, 8} = {4, 5, 8},
KA(6) = {6, 7, 8} ∩ {2, 4, 6, 8} ∩ {1, 3, 6, 8} = {6, 8},
KA(7) = {6, 7, 8} ∩ {3, 7} ∩ {2, 4, 5, 7, 8} = {7}, and
KA(8) = {2, 4, 6, 8}.
Characteristic set KB(x) may be interpreted as the smallest set of cases that are indistinguishable
from x using all attributes from B and using a given interpretation of missing attribute values.
Thus, KA(x) is the set of all cases that cannot be distinguished from x using all attributes. The
characteristic relation R(B) is a relation on U defined for x, y є U as follows (x, y) є R(B) if and
only if y є KB(x).The characteristic relation R(B) is reflexive but—in general—does not need to
be symmetric or transitive. Also, the characteristic relation R(B) is known if characteristic sets
KB(x) for all x є U are known.
In our example(A) = {(1, 1), (1, 8), (2, 2), (2, 8), (3, 3), (4, 4), (4, 8), (5, 4), (5, 5), (5, 8), (6,6), (6,
8), (7, 7), (8, 2), (8, 4), (8, 6), (8, 8)}.
In our example presented in Table 1 let us say that B = A. Then the singleton A lower and A-
upper approximations of the two concepts: {1, 2, 4, 8} and {3, 5, 6,7} are:
A{1, 2, 4, 8} = {1, 2, 4}, A{3, 5, 6, 7} = {3, 7},
A {1, 2, 4, 8} = {1, 2, 4, 5, 6, 8},A {3, 5, 6, 7} = {3, 5, 6, 7, 8}.

30
Note that A{1, 2, 4, 8} = {1, 2, 4}. But the set {1, 2, 4} is not A-globally definable. Furthermore,
the set {1, 2, 4} is not even A-locally-definable, so no set of rules can cover precisely this set.
Similarly, A {3, 5, 6, 7} = {3, 5, 6, 7, 8}, and the set {3, 5, 6, 7, 8} is also not A-locally-definable.
Therefore, in general, singleton approximations should not be used for data mining. The second
method of defining lower and upper approximations for complete decision tables uses another
idea: lower and upper approximations are unions of elementary sets, subsets of U (Grzymala-
Busse 2003). Therefore lower and upper approximations[15][16] for incomplete decision tables
by analogy with the second definition of approximations is defined for completely specified data,
using characteristic sets instead of elementary sets.
There are two ways to do this. Using the first way, a subset B-lower approximation of X is
defined as follows:
BX=U { KB(X) | XєU, KB(X) X }
A subset B-upper approximation of X is
BX=U { KB(X) | XєU, KB(X)∩Xɸ }
Since any characteristic relation R(B) is reflexive, for any concept X, singleton B lower and B-
upper approximations of X are subsets of subset B-lower and B upper approximations of X,
respectively. The second possibility is to modify the subset definition of lower and upper
approximation by replacing the universe U from the subset definition by a concept X. A concept
B-lower approximation of the concept X is defined as follows:BX=U { KB(X) | XєU, KB(X) X }
Obviously, the subset B-lower approximation of X is the same set as the concept B lower
approximation of X. A concept B-upper approximation of the concept X is defined as follows:
BX=U { KB(X) | xєX, KB(x)∩Xɸ }= U { KB(X) | xєX}.
The concept B-upper approximations of X are subsets of the subset B-upper approximations of X.
For the decision presented in Table 7.3, the concept A-lower and A-upper approximations
are:A{1, 2, 4, 8} = {1, 2, 4, 8}, A{3, 5, 6, 7} = {3, 7}, A {1, 2, 4, 8} = {1, 2, 4, 6, 8},A {3, 5, 6,
7} = {3, 4, 5, 6, 7, 8}.For complete decision tables, all three definitions of lower approximations,
singleton, subset and concept, coalesce to the same definition. Also, for complete decision tables,
all three definitions of upper approximations coalesce to the same definition. This is not true for
incomplete decision tables, as our example shows. Singleton B-lower and B-upper
approximations of the set X are subsets of the subset B-lower and B-upper approximations of X,
respectively. The subset Blower approximation of X is the same set as the concept B-lower
approximation of X. The concept B-upper approximation of X is a subset of the subset B-upper
approximation of X. The concept B-upper approximations are the smallest B definable sets
containing X. Rules may be framed by using concept approximations which will have an another
way of research avenue in the context computing.

31
Figure.4 Rule set
3. RESULTS AND DISCUSSION
3.1 Performance Analysis – Fuzzy Vs Rough set Based
This section evaluates the DYCOAD context model approach with the help of both fuzzy sets and
rough sets. Previous section completely describes the roughest based approach. Similarly for
fuzzy sets, Let Chest pain, Breathing Discomfort and Sweating are the separate fuzzy sets. Here
we are making use of CAD (coronary artery disease) Dataset attributes. Then based on the if
then rules of fuzzy set, if (Chest pain = high) and (Breathing Discomfort = yes) and
(Sweating = yes) then patient = Heart Disease
Table 2. Comparative study Fuzzy Vs Rough set
if any one of the data is missing then it is said to be false condition. Hence the comparative study
for fuzzy and roughest is shown Table2.All the entries of the Table2 is represented in
Percentages.

32
Figure 5 Performance analyses – Fuzzy Vs Rough set theory
4. CONCLUSION
In this paper, the context model is designed for CAD (coronary artery disease) dataset using a
context-aware system .Knowledge inference engine is also proposed for context-aware
applications. The reasoning performance ability of our approach showed that context modeling
and development of inference rules are applicable for the wearable computing environment.
The above DYCOAD architecture improves the quality of the service discovery by providing
specific node values and multiple inheritances. The rules are generated using roughest approach
and it is proved that roughest will provide a good clarity around 90% over the incomplete data set
compared to fuzzy set. Thus, the discussed issues and challenges in context aware service
discovery platforms are alleviated with the help of the define DYCOAD approach, Rough set and
CAD Dataset.
REFERENCES
[1] W. Pedrycz, R. R. Gudwin, and F. A. C. Gomide, Nonlinear context adaptation in the calibration of
fuzzy sets, Fuzzy Sets Syst., vol. 88,no. 1, pp. 91-97, 1991.
[2] Dey, A. K. Understanding and using context. Personal and Ubiquitous Computing, 5(1):pp. 4.7. 2001.
[3] Schmidt, a., Beigl, M., Gellersen, H. W., There is More to Context than Location, Computer
Graphics Journal, vol. 23, no. 6, pp. 893-902, Dec. 1999.
[4] W. Sitou and B. Spanfelner, “Towards requirements engineering for context adaptive systems.” in
Proceedings of the 31st Annual International Computer Software and Applications Conference
(COMPSAC 2007), vol. 2. Beijing, China: IEEE Computer Society, 2007, pp. 593–600.
[5] H. Lieberman et al. “Out of context: Computer systems that adapt to, and learn from,context”, IBM
system journal vol. 39,pp345-348,December 2000.
[6] D. Zhang, Z. Yu, and C.Y. Chin, “Context-aware infrastructure for personalized HealthCare,”
International Workshop on Personalized Health, IOS Press, pp.154–163,December 2004.
[7] Strang T., Linnhoff-Popien C., A context modelling survey workshop on advanced context
modeling, reasoning and management as part of ubicomp 2004, the 6th intI. conf. on ubiquitous
computing, pp 33-40, sept 2004.

33
[8] Indulska 1., Robinson R., Rakotonirainy A., Henricksen K. Experiencesin Using CC/PP in Context-
Aware Systems. In Proc. 4th Intl.Conf. on Mobile Data Management, pp. 247-261, Melbourne,
Australia, 2003.
[9] Dey A.K, Abowd G.D, Salber D. A conceptual framework and a toolkit for supporting the rapid
prototyping of context-aware applications. Human-Computer Interaction vol. 16, 2001, p97-166.
[10] McCarty 1. Notes on formalizing contexts. In Proc. of the 13 international joint conference on
artificial intelligence, Bajcsy Ed.M.Kaufmann, pp. 555-560 , California, 1993.
[11] Dey, A., “Providing Architectural Support for Building Context-Aware Applications”, Ph. D. Thesis
Dissertation, College of Computing, Georgia Tech, December 2000.
[12] P. Anandhakumar, G. Nalini priya, L. Priya, Ontology-Based Service Discovery Model For
Distributed Pervasive Healthcare environment“International Conference on Web Intelligence Systems
ICWIS09 pp 87 -90.January 2009.
[13] Z. Pawlak, “Rough Sets”, International Journal of Information and Computer Science, 1982, pp. 341-
356.
[14] Turker Ince, Serkan Kiranyaz, Jenni Pulkkinen, Moncef Gabbouj, Evaluation of global and local
training techniques over feed-forward neural network architecture spaces for computer-aided medical
diagnosis, Elsevier - Expert Systems with Applications, Volume 37, issue 12, December 2010.
[15] Resul Das, Ibrahim Turkoglu, Abdulkadir Sengur, Effective diagnosis of heart disease through
neural networks ensembles, Elsevier - Expert Systems with Applications, Volume 36, Number 4,
May 2009.
[16] G.Alvarez and E.Coiera, “Interuptive communication patterns in the Intensive Care Unit Ward
Round”, Journal .Medical Informatics, Vol.74,no.10, pp779-781, 2005.
[17] G.NaliniPriya, P.AnandhaKumar”Neural Network Based Efficient Knowledge Discovery in Hospital
Databases Using RFID Technology”, Proceeding of International IEEE Conference Tencon,
November 2008.
[18] G.NaliniPriya, P.AnandhaKumar “Efficient Knowledge Discovery in smart environments- A Trend
analysis”, Proceeding of National Conference Tech2011 February 2011.
[19] G.NaliniPriya, P.AnandhaKumar” Object Oriented Database Approach for Context Aware Services in
Distributed Pervasive Healthcare Environment”, Proceeding of National Conference NCACA,
February 2009.

Dynamic Context Adaptation for Diagnosing the Heart Disease in Healthcare Environment Using Optimized Rough Set Approach

More Related Content

Similar to Dynamic Context Adaptation for Diagnosing the Heart Disease in Healthcare Environment Using Optimized Rough Set Approach (20)

Recently uploaded (20)

Dynamic Context Adaptation for Diagnosing the Heart Disease in Healthcare Environment Using Optimized Rough Set Approach