Feng Shui Garden Adviser System (FengShuiGAS)

Shuzlina Abdul Rahmana , Sofianita Mutalibb and Azlinah Mohamedc

Intelligent Systems Special Interest Group,
Department of Systems Science,
Faculty of Information Technology and Quantitative Sciences,
Universiti Teknologi MARA, 40450 Shah Alam, Selangor
Tel: +60-3-5543-5399, Fax: +6-03-5543-5501, E-mail:{ ashuzlina, bsofi, c azlinah}@tmsk.uitm.edu.my

ABSTRACT to develop an adaptive expert system prototype that

provides a corresponding garden design style that
This paper explores an approach to building an satisfies the clients based on their Feng Shui
adaptive expert system prototype in an environment of endeavours.
human-computer collaboration. Components of an
adaptive system are identified, with an emphasis on the 2.0 FENG SHUI OVERVIEW
mechanisms that enable adaptive behavior to occur. An
adaptive expert system is necessary in order to Feng Shui is not related to any religion or belief
communicate with the user and also adapts to user’s system. It is a scientific discipline based on the analysis
needs. The adaptive expert system in this particular of energy (Too, 2002). Feng Shui is not a superstition
project is implemented on a Feng Shui garden design or a set of dos and don’ts. It is the art and science of
domain. A frame -based data representation and rule- understanding the forces of nature in order to design
based approach is applied to this project. In this houses and workplaces that blend with the environment
research, the Feng Shui aspiration is adapted to users’ instead of clashing with it. It aims to help us live in
assessment and choice based on their preferences. This harmony with the world by promoting the flow of
experimental expert system prototype displays low level positive energy and neutralizing or avoiding negative
adaptive capabilities that show sufficient promise to or destructive energy.
warrant further research.
Based on an interview with Master Larry Sang from
Keywords The American Feng Shui Institute, Feng Shui, means
wind and water. It is an ancient Chinese study of the
Adaptive Expert System, Feng Shui, Frame-based, natural environment. Feng Shui can determine the best
Rule-based or most favorable location for anyone and assist that
person to avoid the worst or least desirable location in
1.0 INTRODUCTION any environment. Feng means "wind," and shui means
"water". In Chinese culture, gentle wind and smooth
Nowadays, many people are concerned about their water have always been associated with a good harvest
lawn or garden decorations. The interests of designing and good health, while harsh winds and stagnant water
Feng Shui garden among Malaysian Chinese are have been linked to famine and disease. Therefore,
arising due to the new trendy and healthy lifestyle. The "good" Feng Shui has come to mean good livelihood
service providers for garden design also increases in and fortune, and "bad" Feng Shui has come to mean
numbers day by day. The initial individual who are hardship and misfortune.
interested in Feng Shui garden are confronted with the
problem to design their garden based on their specific Feng Shui basic principles are mainly based on yin and
endeavors that associates with their garden entrance’s yang, the five elements, east or west compass
wind direction. Most of them do not want to hire the directions, the solar system and the environment (Too,
garden designer or Feng Shui expert because the fees 2002). Feng Shui is based upon a set of theories and
are usually too expensive. The landscape service complex calculations derived from the Yi Jing. This
providers on the other hand confronted with the includes an in-depth understanding and application of
problem of understanding the client’s needs and the theories of Yin and Yang, along with the balancing
specifications. They are also confronted with the of the Five Elements (Fire, Earth, Metal, Water, and
problems of recommending landscape style and design Wood) and taking into account the physical
based on the client’s endeavors that related with the relationship between the natural environment and the
Feng Shui. Currently there is no such system that magnetic field.
facilitates both client and the landscape service
provider. Therefore the idea of developing a prototype
system of Feng Shui garden design adviser is to
recommend the best option for both client and
landscape service providers. The aim for this project is

570
3.0 METHODOLOGY reading, elements, flower, furniture, path, and cure. A
set of twenty-four questions was constructed that is
There are three main function operates in FengShui based on recommendation by the expert and reviews.
Garden Adviser System or FenghuiGAS. The first Table 1 shows the questions and the confidence of each
function is an analyzing function whereby the system question in each aspiration. The table also shows the
searches the knowledge based in order to get the expected answer for each aspiration. Each match
advice. The second and third function is the searching answer will carry a weightage that is calculated as
and adapting function respectively. follows:
Percentage of aspiration_value = (0.01/x) * total_correct_match
3.1 Data Analysis (1)
where x is total_expected_answers
The knowledge was acquired from a known landscape
expert and Feng Shui Master while the secondary data User’s answers will be analyzed to determine the
was obtained from literature reviews. Process of weakest Feng Shui aspiration. The weakest Feng Shui
analyzing includes filtering, sorting and data aspiration value is obtained from the lowest aspiration
classification. Knowledge is then represented in case value gained from the above equation.
frames based on aspirations.

Table 1: The question set with the expected answers

Reputation
Knowledge

Prosperity
Creativity
No Question

Wealth
Career

Health

Love
1. There is a mirror in my entry way Y Y Y Y Y

2. I do satisfy with my job Y Y Y

3. I do satisfy by the way my superior treated me Y Y Y

Do you feel that your colleague or friend(s) is/are

4. N N N
backstabbing you?

5. I have a healthy happy pet Y Y Y

6. My home has dusty corner and cobwebs N N N

22 I have fruity plant(s)/tree(s) in my garden Y Y Y Y Y Y

My potted plan(s) is(are) in earth ware/ceramic

23 Y Y Y
container(s)

24 I have musical instruments in my house Y Y Y Y

Total (x) 14 14 12 12 13 15 14 12

3.2 Searching Function

There are eight aspirations namely career, creativity, The lowest aspiration value that was calculated from
health, knowledge, loves, prosperity, reputation and the analysing function will be used as the criteria of
wealth. Each frame has attributes on direction, compass searching. There are seven elements involves in Feng

571
Shui’s suggestion namely, direction, element, compass debugging. If a human expert comes to an incorrect
reading, flower’s colours and types, pathway styles, conclusion, he may be able to learn from the mistake
type and colour of furniture and cures. FenghuiGAS and avoid making the same or similar mistakes in the
searches the database to find those elements that match future. Once an expert system is found to have an error,
with the lowest aspiration value. Figure 1 shows an the only way to correct that error is to reprogram the
example of searching result. One can see that the expert system. In other words, most current expert
weakest aspiration is ‘health’ which is 30.029% and systems are lacking an adaptive capability. Computer
followed by the Garden Design Advice. based adaptive capabilities are essential in situations
where environments change, in situations where
standards of expertise are changing, and in situations
3.3 Adapting Function where there are no historical data and learning occurs
as a task that has been performed.
The adapting function allows the user to change some
According to Pan (2002), nature adaptation is
of the elements value. Feng Shui’s principal only
sometimes referred to as ‘evolution’ whereas ni the
allows two out of seven elements to be changed, which
context of a system it may be referred to as a ‘self-
is type of flower and furniture’s color. For example, the
organization capability’, and, in the case of human
initial suggestion as in Figure 1 for flower is blue and
beings, as a ‘learning capability’. Therefore, since the
furniture type is wood. User may change the flower
intent of expert systems is to emulate human experts,
type from the suggested options whether lobelia,
this paper refers to an adaptive expert system as a
verbena or violet. The same concepts apply to furniture
learning system or, a learner. According to Langley
color whereby the user may change it into other colors.
(1998), learning is any relatively permanent change in
behaviour brought about through experience that is,
through interaction with the environment. An approach
Query (In term of a set of question):
Answers entered by user. to learning, therefore, is through the interaction
between the learner and its environment. The
Comparison – Provided by the system: interaction operates in terms of three basic elements
The weakest Feng Shui (health – 30.029%) that are precepts; actions; and, goals (Shen, 1994).
Thus, a learning process is a process in which a learner
Garden Design Advice (from database):
1) direction: east builds models of its environment.
2) compass: 67.5 – 112.5
3) element: wood Adaptivity is the possibility for the user to change
4) flower: f5 (colour: blue; various aspects of a system explicitly. Adaptive
name: lobelia,verbena,violet)
5) furniture: r5 (based: wood; colour: white) systems may adjust the interface according to the user’s
6) path: p4 (type: bluestone paving; skills, knowledge and preferences. For example, an
shapes: undulating line; adaptable system allows the user to control these
edgings: pressed metal) adjustments, such as user can configure and customize
7) cure: c7 (bagua: c7; five_element: c7)
the user interface as per her or his needs by changing
the properties like colour or font size.

Figure 1: The searching and analyzing function 4.2 System Framework

FenghuiGAS was developed by adapting to Pan (2000)

4.0 ADAPTIVE EXPERT SYSTEMS architecture. According to him, ni terms of inference
process, an adaptive expert system describes a
generate-and-test loop that endows a system with a
An expert system could be viewed as one heuristic rule,
learning capability, while a conventional expert system
or if-then statement, since all rules in the system
describes a unidirectional flow with no adaptive
working together to define a specific condition, which
behaviour. From a system structure perspective, an
is achieved by asking user relevant questions.
adaptive expert system differs from a conventional
Thereafter, the system reacts with one or more actions.
expert system as follows;
It is postulated that an expert system becomes adaptive
during an interaction with a human user if this system • Unlike a static knowledge base in a conventional
proves to be capable of the formulation of new relevant expert system, an adaptive expert system contains
questions and the adjustment of actions (Pan, 2000). a dynamic knowledge base. Knowledge is
represented by a knowledge network (such as
4.1 Adaptive Approach objects and their interactions), that can be
dynamically modified in real-time based on
A significant amount of time is required in building an interactions with the human user.
expert system with many hours are on testing and

572
• In addition to containing the capabilities of a determines user’s aspirations and advises on garden
conventional inference engine, an adaptive expert design based on the questions and answers session.
system is capable of real-time adaptation to
feedback and restructuring of the knowledge On the other hand, if users dislike the advice, they may
network. change the preferences. In implementing this situation,
FengShuiGAS will apply the backward mechanism.
The framework of FenghuiGAS consists of two major The adaptive function will prove the goal using the
components: an application interface and an adaptive facts and rules whereby the aspiration value acts as a
module. The application interface is divided into goal state. The modification can be done until users are
inference engine and knowledge network. The satisfied with the recommendation.
inference engine infers through condition-satisfaction
action selection. In the rule-based system, it searches 5.0 DISCUSSION
for the rules satisfied by facts or objects, and executes
the rule with the highest priority. The knowledge base In the first stage, the knowledge is represented using a
is composed of knowledge nodes and their connections. frame -based as it provides a natural way for the
Meanwhile, the adaptive module performs the structured and concised representation of knowledge
recommendation task. The mechanism provided by this (Higginbotham 1996, Nebel, 1999). It can organize
system demonstrates a primitive adaptive capability of knowledge in slots in order to describe various
a computer-based expert system as shown in Figure 2. attributes and characteristics of the object in a
meaningful way. The system is developed using object-
oriented programming technique as it defines the object
Adaptive Module in a real world through a natural way of representation
into computer environment.

Five sets of frame-based knowledge representations

were constructed which was based on Feng Shui’s
elements. Figure 3 shows an example of frame -based
knowledge representation that has been constructed.

During the development process, we have found that it

is difficult to constantly represent the knowledge in the
frame -based because it has to deal with many
constraints. For example, the system required the user
to answer a set of multiple choices questions (in the
form of Yes/No). It is hard to interrelate between the
outputs and the system assessment with the frames.
Therefore, frame representation was used only to
represent the information in a real world.

The system used rule-based technique that combined

4.3 Inference Engine with a weighted score for each primary value. The
results that are produced by the system are the garden
Inference engine is the heart of rule-based design style including the flower’s colours and names,
implementations where it provides the mechanism that the path styles, auspicious furniture based and colours
controls the course of consultation, combining the rules and Feng Shui’s cures that are suitable with the user’s
in the knowledge base with input data to develop weakest aspiration. If the user does not satisfy with the
recommendation. Both forward chaining and backward result or the system advice, the user can change the
chaining mechanisms were used in this project. flower types and the furniture colour.
FengShuiGAS evaluates the aspiration once the users
answered all the 24 questions. Each answer will carry a
weight for each aspiration and then the percentage of
each aspiration is calculated.

Forward chaining and backward chaining were used as

a state space search. Forward chaining or data driven
reasoning, works by starting with facts and rules, and
tries to generate new facts. In FengShuiGAS, this
mechanism was applied for analyzing and searching
function. Forward chaining is used when the system

573
 CLASS : INSTANCE :

Figure 3 : An example of frame-based data representation

The explanation facility reacts when the system does will be benefited to both clients and landscape provider
not produces result or advises that satisfies with user who are interested in a Feng Shui’s garden design.
needs. In addition, the explanation facility was created
as to give more explanations about how and why user By using the weakest aspiration generated from the
gets the result. If the user click the button “Why I get system, both clients and landscape service provider will
this result?” on the result page, the system will prompt be facilitated on how to design the garden, considering
a reason page to inform the user on how the system elements of Feng Shui. It is hoped that clients’ budget
evaluated the result. Meanwhile, the explanation of hiring expert garden designer can be saved. The
facility can be viewed when the user click the button prototype may encourage and expand the use of e-
“No – Changed” in the adviser form. For example, if commerce application in a landscape industry. Further
the user clicks the button on the flower tips section, the improvement may be obtained by incorporating with
system will display the message “Flower: You cannot other domain, such as home and office interior design
change the colour because this is the auspicious colour and the use of 3D environment interface design for
to cure your weakest Feng Shui. This is to make sure interactivity.
the user understand why such result or advice is given
to them. Acknowledgment

The authors would like to sincerely thank to our

6.0 CONCLUSION research student, Norhayati Salehudin for her valuable
work that support this research.
This paper provides an alternative approach by using
frame -based data representation and rule-based to
develop a prototype of expert system to advise on Feng
Shui aspirations. If FengShuiGAS is commercialised, it

574
REFERENCES Roberts, S. (2005). Feng shui bagua (ba gua) -
Beattie, A. (2003). Feng shui garden design: Creating excerpted from fast feng shui: 9 simple principles
serenity. Singapore: Periplus Edition. for transforming your life by energizing your
Buzzanca, G. (1998). A rule based expert system for home. Retrieved June 27, 2005, from
musical style recognition. USA: Harvard http://www.fastfengshui.com/feng_shui_bagua.ht
University, Cambridge. Retrieved February 20, m
2005 from Shen, W. (1994). Autonomous Learning from the
http://citeseer.ist.psu.edu/cache/papers/cs/27677/ht Environment. New York: W. H. Freeman and
tp:zSzzSzwww.conservatoriopiccinni.itzSz~g.buzz Company. Publisher: Springer Science+Business
ancazSzA_rule_based_expert_system.pdf/buzzanc Media B.V., Formerly Kluwer Academic
a01rulebased.pdf Publishers B.V. 143 – 165. Retrieved February 27,
Chien, S. (1993). A explanation facility for rule -based 2005, from
expert systems. San Luis Obispo: MS Thesis in http://www.springerlink.com/app/home/contributio
Architecture, Cal Poly, San Luis Obispo, CA. n.asp
Choi, J. (2002). A rule -based expert system using an Stewart, M. K. (2004). The eight aspirations formula.
interactive question-and-answer sequence. Boston: Feng Shui Bostan Inc. Retrieved
Retrieved February 27, 2005, from September 20, 2005, from
http://www.cobblestoneconcepts.com/ http://www.fengshuiboston.com/8aspirations.html
ucgis2summer2002/choi/choi.htm Tickel.com: How good is your feng shui online test?
Cohn, P. (March 2000). Feng shui in the garden. Retrieved October 10, 2005 from
Retrieved October 27, 2005, from http://web.tickle.com/tests/fengshui/paidresult.jsp?
http://www.findarticles.com/p/articles/mi_m1082/i c=42334&where=paidresult&test=fengshuiogt
s_2_45/ai_60021668 Too, L. (2002). Complete illustrated guide to feng shui
Dale, N. Weems, C., & Headington, M., (1996). for gardeners. London: Element.
Programming and Problem Solving with Turban, E. (1992). Expert system and applied artificial
C++.Boston: Jones and Bartlett. intelligence. New York: Macmillan Publishing
Durkin, J. (1994). Expert System: Design and Company.
Development. Macmillan Publishing Company,
New York.
Langley, P. (1998). Elements of machine learning. San
Francisco: Morgan Kaufmann.
Encyclopedia Britannica. (2005). Retrieved on
March 10, 2005, from Encyclopædia Britannica
Premium Service -
http://www.britannica.com/eb/article?tocId=91097
61
FAQ. The American Feng Shui Institute. (2004).
Retrieved August 15, 2005, from
http://www.amfengshui.com/faq.htm
Higginbotham, D. J., Moulton, B. J., Lesher, G. W.,
Wilkins, D. P. & Cornish, J. (1996). Frametalker:
development of a
frame -based communication system. Retrieved June
25, 2005, from
http://www.enkidu.net/downloads/papers/HiMoLe
WiCo00a.pdf.
Johnson-Laired, P., (1993). Human and machine
thinking. Hillsdale, New Jersey: Lawrence
Erlbaum Associates.
Nebel, B. (1999). Frame -based systems.
Massachusetts: Massachusetts Institute of
Technology. Retrieved July 10, 2005, from
http://www.cs.umbc.edu/771/current/papers/nebel.
html
Pan, X., (2000). An Experimental Adaptive Expert
System. Collaborative Agent Design (CAD)
Research Center, USA: Cal Poly, San Luis Obispo,
CA. Retrieved February 20, 2005, from
http://www.cadrc.calpoly.edu/pdf/adaptive_expert.
pdf

575