There are a number of unique plant types in the world which are on the brink of extinction, or are so extremely limited in range. For example the beautiful palms such as the Talipot Palm (Corypha elata), Johannis teysmannia in Sarawak and Malaya, and also Livistona in West Sumatra. The largest flower in the world, Rafflesia arnoldii and related species, are going to vanish unless more nature reserves are created for such remarkable plants in Sumatra, Borneo, Java, and the Philippines.

Rafflesia is a genus of flowering plants that is made up of 23 known species. The best known of these species is Rafflesia arboldii which has the distinction of being the world's largest flower, reaching a diameter of about three feet. The genus Rafflesia gets its name from Sir Stamford Raffles, the founder of the British colony of Singapore (Walter et al., 1998).

Rafflesia is a unique plant because of its dimensions unlike other flowers. This circumstance makes it widely known. It is also become an icon for conservation especially of the rain forest area. The rarity creates interest among nature lovers, tourists and the general public. The fact that some species may be in the brink of extinction alarms conservation groups. Its little-known biology and reproductive ecology spurs the interest of botanists and ecologist (Nais, 2004). Rafflesia is also thought to be one of the rarest of all plant genera which is only found in Borneo, Sumatra, Java, Peninsular Malaysia, Thailand and the Philippines.


Malaysia is very fortunate for being one of the habitats of Rafflesia. But unfortunately, all of the known species of Rafflesia are threatened or endangered. In Malaysia the Rafflesia is only a "Totally Protected Plant" by law in Sarawak. In Sabah and Peninsular Malaysia it is only safeguarded by laws when found in protected areas like National or State Parks. In 2002, 44 out of the 83 Rafflesia flowers found in Sabah were outside of designated conservation places (Sabah Travel Guide, 2004).

Eight out of the 23 known species of Rafflesia can be found in Malaysia, most of them in the jungles on the island of Borneo. Some species of Rafflesia are endemic species. That means these species are native and can be found only in that location. For example, Rafflesia tengku-adlinii seems to be endemic to Sabah only while Rafflesia tuan-mudae endemic to Sarawak only. Because the Rafflesia is only found in specific areas and its locations often difficult to reach, and because it only blooms for a very short time, its life cycle or the methods of pollination and seed dispersal is very unclear. This makes the appropriate methods to conserve it quite difficult to be found.

Other than that, in Peninsular Malaysia flower buds are still sold as traditional medicine. The buds are seen as a sign of fertility, and are given to help mothers recover after birth. The over collection of these buds has not helped with conservation efforts but further drastically reduced the number of Rafflesia in the wild, accentuating the problem the alarmingly fast transformation of jungles into palm oil plantation creates (Sabah Travel Guide, 2004). The Rafflesia is a delicate plant that relies on an intact environment and as such is naturally extremely vulnerable to deforestation and development.

Conservation must be done to protect this species from extinct. The expert system may help the user to identify the Rafflesia species in Malaysia. By using an expert system, the user may gather information about approaches to conserve the Rafflesia.


There are some purposes for this research. The main objectives of this research are:

  • to identify the Rafflesia species in Malaysia based on their physical characteristics.
  • to develop an expert system which help the public to recognize the Rafflesia species in Malaysia.
  • to verify system performance in order to make it applicable to the real world.


The scope of this research is mainly about the Rafflesia flower and the approaches to conserve it. This research also about the system named an expert system which using Macromedia Dreamweaver 8. The system is the tool or mechanism which contains all collected information, recommendation and opinion from many expertises and also results from so many researches done. The target users of this system are tourists, publics, nature lovers and also those who interested in conserving the Rafflesia plant.

The system that will be developed will provide the public about the Rafflesia species in Malaysia as well as the approaches to conserve the Rafflesia. By using an expert system, it would easier the public to access about this endangered plant.


This thesis consist of five chapters; introduction, literature review, methodology, results and discussion and conclusion.

Chapter 2 is about literature review. This chapter includes the Rafflesia characteristics, the diversity and habitat of Rafflesia, identification of Rafflesia species as well as treats and conservation of Rafflesia. Introduction of expert system also included in this chapter. Expert systems typically have three basic components; a knowledge base, an inference engine and user interface.

Chapter 3 is about methodology. This chapter contains the development stage of expert system for Rafflesia species identification. There are five stages in developing an expert system which are task analysis, knowledge acquisition process, prototype development, expansion and refinement and lastly verification and validation.

Chapter 4 is about the result and discussion. This chapter consists of the architecture of the system and also the flow in the developing system process. This research is using an IF-THEN rule in form of asking question to the user.

Chapter 5 is about the conclusion. This chapter consists of conclusion for overall of this research. It includes the expert system technology, the prototype development of expert system and the recommendation to make the system move effective and also ways to improve it.




The Rafflesia is one of the most magnificent flowers ever known to the botanical world. It is such a big flower with odd appearance, exceptional, rare and also mysterious. It is also immense scientific and public interest. Rafflesia in bloom has been described as simply awesome (Nais, 2001). This chapter will discuss about Rafflesia characteristics, diversity, habitat, species identification as well as treats and conservation of Rafflesia.


In general, Rafflesia flowers consist of five leathery petals that are orange in colour and mottled with cream-coloured warts (Attenborough, 1995). The flower also has no leaves, stem or roots.

The dramatic Rafflesia flowers are the largest single flowers in the world; the leathery petals can reach over 90 centimetres across (Attenborough, 1995). Rafflesia is a parasite that depends completely upon its host which supplies nutrients and water to the flower. These host plants are vines of Tetrastigma spp., and the Rafflesia plant is itself not visible until the reproduction stage when flowers first bud through the woody vine and then open into the magnificent spectacle that is world-renowned today (Nais, 2001).

Most flowers in the genus give off and smell of rotting flesh, hence the local called it the "corpse flower". When in bloom, the flower displays its five fleshy 'petals' or so called perigone lobes. The diameter of the various species of Rafflesia flower ranges from approximately 20 cm to a record diameter of 106.7 cm (Meijer, 1985). Other than that, the unique part of this giant flower is the flowers can take up to ten months to develop from the first visible bud to the open bloom, which may last from 5 to 7 days only.

Currently 17 species of Rafflesia are recognised and these mainly differ in the morphology of their flowers (Nais, 2001). There is a deep well in the centre of the flower containing a central raised disc raised that supports many vertical spines (Attenborough, 1995). The sexual organs are located beneath the rim of the disk, and male and female flowers are separate (Attenborough, 1995).


There are 23 completely known species and 4 incompletely known species of Rafflesia as recognized by Meijer on 1997. Table 3.1 shows the known species of Rafflesia and Table 3.2 shows the unknown species of Rafflesia as recognized by Meijer on 1997.


Rafflesia is restricted to the western part of the phytogeographical region of Malesia, which is known as the Sunda shelf (Nais, 2001). The region is floristically distinct, with a clear boundary from surrounding region (Steenis, 1950). Table 2.3 shows the genus distribution in their landmass location. Number in parentheses denotes the number of Rafflesia species present in each area.

Book of Rafflesia Magnificent Flower of Sabah by Kamarudin Mat Saleh (1991)

A total of 23 names of Rafflesia species have been published between 1821 and 1988 (index Kewensis, 1994). Six of the names are now considered synonyms, and a further four have inadequate type material and are treated as insufficiently known species (Meijer, 1997).

The western most extension of Rafflesia is Acheh District, Sumatra, followed by the Ranong Province in Thailand, about 5 km from the Myanmar border (Meijer & Elliot, 1990; Banziger, 1991; Elliot, 1991; Banziger et al., 1993). The eastern limit is Mount Apo Timur and at Gunung Dadum in Eastern Sabah (Nais, 2001). The northern most limits are at Mount Makiling, Los Banos Province, on Luzon Island in the Philippnes (Madulid & Agoo, 1996), and the southern limit is the province of Java, Indonesia. The distribution of Rafflesia's genus is shown in Figure 2.1.


The first description of the morphology of Rafflesia was made by Robert Brown (1821, 1835), who provided a detailed and meticulous description and illustrations of the male and female flowers of Rafflesia arnoldii (Nais, 2001). The taxonomy of Rafflesia is based entirely on the floral morphology (form and structure) of the flower with most emphasis on the outer appearance. The current species delimitation of Rafflesia is based on eight major characters. These eight variable characters are listed below (Nais, 2001) and the radial section of Rafflesia flower drawn by Yong Ket Hyun, after Meijer 1985 are shown in Figure 2.2.

  • Size (diameter of open flowers varies from 15 cm in R. manillana to nearly 1 m in R. arnoldii);
  • Diameter of the diaphragm aperture (ranging form 3-9 cm in R. micropylora, to about 20 cm in R. arnoldii);
  • Number of disk processes (from none in R. rochussenii to 20-60 in R. arnoldii);
  • Size and number of white spots (called blots, specks or warts) on the perigone lobes and diaphragm (from few to large in R. hasseltii to numerous and small in R. arnoldii);
  • Number and size of the 'windows' on the inside or lower surface of the diaphragm (3-5 rings of round spots in R. micropylora, seven rings of round spots in R. kerrii, or five rings of elongate oval spots in R. pricei)
  • Number of anthers (from about 15 in R. manillana to 40 in R. arnoldii);
  • Structure and length of ramenta, and position of their occurrence (from short and more or less postulate in R. manillana to up to 12 mm long in R. micropylora; in R. schadenbergiana, ramenta occur on the undersite of the diaphragm; ramenta in various species may have apices branched or unbranched, swollen or crateriform); and
  • Number of annuli at the base of the perigone tube and column, all species has either 1 or 2 (for example, 2 in R. pricei).

In this research, the focus is on the Rafflesia species in Sabah, Sarawak and Peninsular Malaysia only. There are eight species of Rafflesia can be found around Malaysia which is four species (R. azlanii, R. cantleyi, R. hasseltii and R. kerrii) located in Peninsular Malaysia, three species (R. tengku-adlinii, R. keithi and R. pricei) can be found in Sabah and four species (R. tuan-mudae, R. pricei and R. keithii) in Sarawak.

2.5.1 Rafflesia cantleyi Solms-Laubach

Rafflesia cantleyi was named after M. Cantley, curator of the Singapore Botanic Gardens in 1880 to 1886, who collected the type specimen in 1881. This species was described by H. Graft Solms-Laubach based on a collection made by M. Cantley in 1881. Rafflesia cantleyi has open flower dimension from 30 to 55 cm in diameter. Its perigone lobes are up to 14 cm long and 18 cm wide. It has 6 to 8 whitish warts in radial and lateral directions, about 10 in the basal row. The diaphragm opening is 4 to 8 cm across and rounded shape or sometimes angular. It has 5 concentric rings of oval white blots. Its ramenta is 2 cm long in upper type while middle and lower type is 10 to 12 mm longs, or sometimes branched, almost all with swollen apices. Its number of anthers is from 20 to 25 and this species can be found in Peninsular Malaysia only (Perak, Kedah, Perlis, Selangor, Kelantan, Terengganu, Pahang and Tioman Island). Figure 2.3 shows the picture of Rafflesia cantleyi.

2.5.2 Rafflesia hasseltii Suringar

Rafflesia hasseltii was described by Suringer in 1879 from discovery in Central Sumatra. R. hasseltii has an open flower dimension from 35 to 50 cm in diameter. It has 10 to 13 cm long and 14 to 17 cm wide of perigone lobes. It has whitish-pinkish blots across and large size of blots ranging from 5 x 3 to 10 x 1 cm. The clear contrast of snow white blots on bright brick-red background easily distinguishes this species from others. Its window is whitish or pale yellowish with a dark brown zone near the rim and the compound ramenta near the attachment point of the diaphragm gradually become white blots (window) on the lower part of the diaphragm. The R. hasseltii ramenta's upper type is toadstool-like compound ramenta which gradually becoming the white blots of the windows. While for middle and lower type is generally linear with swollen apices. The number of anthers is 20 and this species can be found in Sumatra, Peninsular Malaysia and Borneo and its altitudinal distribution is from 400 to 600 m. Figure 2.4 shows the picture of Rafflesia hasseltii.

2.5.3 Rafflesia keithii Meijer

Rafflesia keithii was named after Harry G. Keith, the former Conservator of Forests in British North Borneo or nowadays called Sabah, Malaysia. This species was described by Willem Meijer in 1984. Rafflesia keithii has an open flower dimension from 80 to 94 cm in diameter. The perigone lobes are 10 to 12 cm across at the median of the lobes but sometimes it has six-lobed. Its colour is numerous white warts with dense, more or less of the same sizes. The diaphragm opening is normally 5 concentric of white warts in about 40 radial rows, each surrounded by a dark red-brown margin. Its window has 5 to 6 rings of large, white blots, those nearer to the rim merging. The ramenta for upper type is 5 to 6 mm long which often fascicled (in bundles) middle and for lower type is solitary and only some with a swollen head. Rafflesia keithii always has 40 numbers of anthers. This species can be found only in Borneo with 250 to 940 m of altitudinal distribution. Figure 2.5 shows the picture of Rafflesia keithii.

2.5.4 Rafflesia kerrii Meijer

Rafflesia kerrii was named after A.F.G. Kerr, Thailand's first Government Botanist, who collected the specimen from which the specimen was described. This species was described by Willem Meijer in 1984. Rafflesia kerrii has an open flower dimension of 50 to 70 cm in diameter. It has 13 to 20 cm long and 19 to 24 cm wide of perigone lobes. Its colour is dull red with brownish tinge and have numerous and scattered warts with 3 to 4 mm space between them. The size of warts for R. kerrii is the smallest compared to other species. The diaphragm opening ia about 12 to 17 cm across and have upper face with 3 to 4 concentric rings of white spots surrounded by a dark red margin. The characteristics for its window are bright white in colour with roundish to elliptic blots and up to 10 mm of diameter. This species of Rafflesia has ramenta that mostly unbranched and only slightly swollen at apex. Its anthers consist of about 26 to 31. The distribution of Rafflesia kerrii is surrounding Thailand and Peninsular Malaysia which at attitude from 500 to 1000 m. Figure 2.6 shows the picture of Rafflesia kerrii.

2.5.5 Rafflesia pricei Meijer

Rafflesia pricei was named after William Price, a honourary plant collector for the Royal Botanic Garden, Kew, who collected this species along the trail to the Mamut Copper Mine. This species of Rafflesia was described by Willem Meijer in 1984. Rafflesia pricei can only be found in Sabah only. It has 16 to 45 cm of opening flower dimension. It has 40 to 80 raised whitish spots (warts) surrounded by brick-red background. The spots range from 1 to 4 cm in length and the surface is minutely rugolose giving a matted appearance while the red background is densely papillose (Beaman at al., 1988). The diaphragm opening is about 5 to 6 cm in diameter and the diaphragm has 4 to 5 irregular rings of white spots smaller than those of the perigone lobes, surrounded by brick-red areas that grade into the cream-white background, the inner edge has a narrow white margin with a continuous reddish-brown area just outside the white rim (Beaman et al., 1988). The R. pricei has 4 to 5 concentric rings of large white window panes or blots and contracting with the bright red background. The interior of the perigone tube from the base of the tube to the lower edge of the diaphragm is lines with wine-red ramenta. The upper type (near the diaphragm's opening) is about 2 to 6 long while middle type (near the insertion of the perigone lobes) is about 4 to 6 cm long and the lower type (near the base of the perigone tube) is about 6 to 7 mm long. The number of anthers for R. pricei is 20. Figure 2.6 shows the picture of Rafflesia pricei.

2.5.6 Rafflesia tengku-adlinii Mat Salleh & Latiff

Rafflesia tengku-adlini was named after Tengku Datuk (Dr.) Adlin Tengku Zainal Abidin, a keen naturalist and conservationist in Sabah who facilitated the documentation of the species after its recovery. This species was described by Kamarudin Mat Salleh and A. Latiff Mohamed from a specimen collected at Mount Trus Madi, Sabah in the year of 1989. R. tengku-adlinii is endemic species to Sabah only at altitude 610 to 800 m. Its opening flower dimension is about 20 to 25 cm diameter while its perigone lobes are 7 to 12 cm long and 12 to 16 cm wide. The colour of R. tengku-adlinii is bright to dull orange throughout with warts throughout the upper surface except near the diaphragm opening. The diaphragm opening is up to 12.5 cm wide and about 3 cm in diameter. It has no windows and the lower diaphragm covered with ramenta. The ramenta occurs right up to the opening of the diaphragm with 3 to 5 cm long, apices swollen, upper, middle and lower types all have fine bristles. The number of anthers for this species is 20. Figure 2.8 shows the picture of Rafflesia tengku-adlinii.

2.5.7 Rafflesia tuan-mudae Beccari

Rafflesia tuan-mudae was named after Mr. Carlo (Charles) Brooke Tuan Muda of Sarawak. This species wasfirst collected from Mount Pueh, Sarawak, from which specimen the species was described by Beccari in 1868. Rafflesia tuan-mudae also is an endemic species. It only can be found in Sarawak. The opening flower dimension can reach from 44 to 92 cm in diameter. The numbers of perigone lobes is usually 5, sometimes 6 (at Gunung Gading National Park, Sarawak (personal observation), or even 7 (in Cagar Alam Gunung Raya Pasi (Zuhud et al., 1998). The colour of R. tuan-mudae is much like R. keithii, but it only has 5 to 8 warts across the median. The diaphragm opening is from 15 to 18 cm and the number of anthers is unknown. Figure 2.9 shows the picture of Rafflesia tuan-mudae.

2.5.8 Rafflesia azlanii Latiff & M. Wong

Rafflesia azlanii is endemic species to Peninsular Malaysia only at altitude 150 to 400 m. Its opening flower dimension is about 38 to 50 cm diameter while its perigone lobes are 9.5 to 10.5 cm long and 12 to 14.5 cm wide. The colour and pattern of R. azlanii is large and continuous (not all like R. hasseltii) whitish warts with brick-red background. The diaphragm opening is from 4.7 to 5.5 cm in diameter. During an early stage of blooming, the open diaphragm is about 7 cm and when the flower bloom fully, the open diaphragm become wider up to 18 cm. The flower's window is large whitish scattered of warts. The ramenta for upper type is 6 mm long while for lower type is 4 mm long. The number of anthers for this species is unknown. Figure 2.10 shows the picture of Rafflesia azlanii.


Rafflesia is one of the most threatened and also one of the rarest plants in the world. The existence of this prodigious flower is precarious and it will eventually become extinct without active conservation efforts.

2.6.1 Threats

Rafflesia are inherently rare as a result of a number of factors of their life cycle; they have a double habitat specialisation, as they can only successfully parasitise particular species and these species in turn are found only in specific habitats (Nais, 2001). In addition to this factor, there is an extremely unbalanced sex ratio in the Rafflesia flowers observed, with many more male than female flowers (Nais, 2001).

Flower buds have a high level of mortality and only 10 to 18 percent go on to bloom, these only lasting for a few days; the chances of a male and female flower being in bloom at the same time in a close enough vicinity to be pollinated is therefore extremely slim (Nais, 2001). In addition to these inherent factors, there is widespread habitat destruction within much of the rain forested area of Southeast Asia and Rafflesia buds are also collected for traditional medicine to treat fertility problems, in parts of their range.

2.6.2 Conservation

The tropical rain forest is the most threatened environment and has experienced the greatest loss of species during our lifetime (Lucas & Synge, 1981). Human disturbance is one of the factors that result great loss of this natural habitats and species. Under the present circumstances, Rafflesia appears to be one of the genera approaching extinction (Nais, 2001). Therefore, the tropical rain forest and all their inhabitants must be the main focus in conservation efforts.

2.6.3 Conservation Status of Rafflesia

The World Conservation Union, IUCN (1984, 1988, and 1997) established five main categories to highlight the conservation status of species:

  • Extinct (no longer known to exist in the wild)
  • Endangered (species that have a high likelihood of becoming extinct in the near future)
  • Vulnerable (species that may become endangered in the near future because populations are decreasing in size throughout the range)
  • Rare (species that have small total numbers of individuals, often due to limited geographical ranges or low population densities)
  • Insufficiently known (species that probably belong in one of the preceding categories but are not sufficiently known to be assigned to a specific category).

The conservation status of Rafflesia can only be reliably assessed by acquiring and analyzing extensive field data of each species. These data include the distribution and the characteristics of sites, rarity and reproductive ecology. The conservation status of all Rafflesia species based on current knowledge which is from published account recognized by IUCN (1997), present analysis using the WCMC/IUCN classification of Conservation Status and present analysis using the IUCN's new Categories of Conservation Status are shown in Table 2.4.

2.6.4 Approaches to Conserve Rafflesia

There are two approaches can be taken according to Nais (2001). One of the approaches and also the best conservation approach for any species is in situ (or on site) conservation which mean leaving it to grow wild in its original habitat. In situ conservation is usually more effective than other approaches because natural condition often impossible to duplicate artificially. In situ conservation strategy of Rafflesia involves:

i) Protection inside established Conservation Areas

One of the major problem to conserve the Rafflesia is because so many tourist eager to pay homepage to existing Rafflesia sites, cause massive trampling, even to level where some populations are trampled to extinction. The mechanisms to control over-visitation are inadequate, and the infrastructure to minimize its impact is not well developed. One way to avoid trampling is by constructing boardwalks or walkways over Rafflesia population. In addition, Rafflesia sites outside conservation areas need to be made available for tourists, thus reducing the pressure for population within conservation areas (Nais, 2001).

ii) Designating New Conservation Areas

It may not be possible to make each Rafflesia site become conservation areas. A more plausible strategy is to synergize the Rafflesia cause with other issues, such as the protection of forests for water catchment, total habitat and biodiversity conservation, and also for nature tourism development (Nais, 2001). This may allow larger area to be conserved to protect their habitat and also their population.

iii) Protection of Sites in other Areas

Many Rafflsia locations are outside the protected areas which within the land belong to the indigenous community. Cooperation from the landlords is very important in order to conserve the Rafflesia. By doing opening sites for tourism is one viable conservation solution and apart from that, it also can generate income for the landowners.

The second approach is by doing ex situ conservation. Ex situ conservation (sometimes referred to as 'off-site' conservation) is the conservation of plants away from their natural occurrence [Given (1994) cited in Nais (2001)]. This approach includes conserving whole plants or plants in botanical garden and gene banks, as well as using laboratory techniques such as tissues culture for their propagation and preservation. Bringing Rafflesia into cultivation has always been appealing. Its cultivation would important not only for its conservation, but also as an important step towards utilization of the plants for ecotourism purposes (Nais, 2001).

Prior to the success by Nais et al. (in press, 2000), many researchers had unsuccessfully tried various methods and techniques of ex situ propagation of Rafflesia. These attempts have included efforts to grow Rafflesia by way of seed insertion into host plants and the translocating of Tetrastigma plants with Rafflesia buds (Nais, 1997; Nais & Wilcock, 1999). Similar cattempts conducted by various other people have also not been successful, for example: seed germination, tissues culture of Rafflesia and grafting of infected host plants into uninfected ones (Ghazally, 1991; Latif & Mat-Salleh, 1991; Zuhud et al., 1998). These failures have initially led some researchers to believe that ex situ cultivation possibly may never be a viable option for the conservation of Rafflesia species (Meijer, 1997) until Nais et al. achieve successful in 1999.


Artificial Intelligence (AI) is a branch of computer science that is principally concerned with using computational models to understand how humans think (Tanimoto, 1987). Major research areas include expert systems, search methods, knowledge representation, logical and probabilistic reasoning, learning, natural language understanding, vision, and robotics (Cohen and Feigenbaum, 1982).

The most successful application of Artificial Intelligence so far is the development of Decision Support System (DSS), particularly expert system, which is a computer program that act as a 'consultant' or 'advisor' to decision makers (Generation5, 2005).

According to Turban & Aronson (2001), an expert system is a system that uses human knowledge captured in a computer to solve problems that ordinarily require human expertise, or, a computer program that can solve problems in a specific area of knowledge (the problem domain) as well as a human expert (O'Keefe et al., 1987), or, that automates tasks that are normally performed by specially trained or talented people (Shannon et al., 1985). Usually when an organization having problem to solve and have to make critical decision, they often turn to consultants or experts seeking for advice. These experts or consultants have specific knowledge and experience in the problem area. They are aware of alternative solutions, chances of success, and costs that the organization may incur if the problem is not solved. Experts can diagnose problems correctly and solve them satisfactorily within a reasonable time frame.However, human experts are expensive, and they may not be readily available.

Expert systems are an attempt to mimic human experts (Turban, Rainer, & Potter, 2001). The expert system begins by asking questions about the problem to be solved. When the needed information has been gathered (inputted by user), the system offers suggestions about how the problem can be solved (McEneaney, 1992). According to Wentworth (1993), expert systems differ from conventional programs in the way they store and use information. In a conventional program, the operations never vary as the programmer predetermines them. The conventional program contains precisely defined logical formulas and data, and if any data element is missing, the program will not run. The expert system, like the human expert, contains heuristic information and can function with incomplete information.

2.7.1 Components of Expert System

An expert system is typically composed of at least three primary components. These are the inference engine, the knowledge base, and the working memory (Wikibooks, 2010). An expert system, also known as knowledge-based system, uses the knowledge and experience of experts to solve problems in a reasonable period of time. Human experts solve problems by using their factual knowledge and reasoning ability. In the other hand, an expert system uses its knowledge base and inference engine to perform a similar task. Figure 2.1 shows the main components in an expert system.

The knowledge base supplies specific facts and rules regarding a domain, while the inference engine offers the reasoning ability that allows the expert system to make conclusions. The user interface is the medium between the expert system and the user. An expert is someone who has the ability to achieve a specific task efficiently by using his or her skills, experience, and knowledge in a specific domain. Knowledge Base

Expert systems are based on human knowledge and reasoning patterns (Wikibooks, 2010). The knowledge base an expert uses is what he learned at school, from colleagues, and from years of experience. The more experience he has, the larger his store of knowledge. Knowledge allows him to interpret the information in his databases to advantage in diagnosis, design, and analysis (Edward and Robert, 1993). This knowledge must be extracted from a human expert by a specialized knowledge engineer. Knowledge engineers ask the expert questions about his knowledge and his reasoning processes, and attempts to translate that into a computer-readable format known as a knowledge base.

The knowledge base grows and expands as the expert behind it adds more knowledge to it. The knowledge base is just like a database which are used to stores information. Some of the ways used to represent knowledge in a knowledge base are scripts (used mostly in natural language systems), logic, processes, rules, frames, and semantic nets (Wikipedia, 2010). The choice of method to present knowledge depends on how the knowledge engineer chooses to think about the knowledge and which representation lends itself most efficiently to retrieval and deduction of facts (McKinion and Lemmon, 1985).

The knowledge base is the module around which the expert system is built. It contains the formal representation of the information provided by the domain expert. This information may be in the form of problem-solving rules, procedures, or data intrinsic to the domain. To incorporate this information into the system, it is necessary to make use of one or more knowledge representation methods (Navin, 1997).

The most common form of knowledge base representation is rule-based. A rule is a conditional statement that specifies an action that is supposed to take place under a certain set of conditions. Rules in an AI program can be somewhat similar to if--then statements in conventional programming languages. However, most conventional programs contain only a relatively small number of possible paths at each step that calls for branching. In contrast, the conditionality embedded in AI problems is so great that the number of paths that can be exploited explodes combinatorial. In conventional programming, the rules are imbedded directly into the program and consequently require considerable effort to develop, debug, and maintain. In a rule-based system the rules are entered into the knowledge base without programming (McKinion and Lemmon, 1985). Inference Engine

The inference engine is where a decision or solution a problem is drawn. The expert system reasons or makes inferences in the same way that a human expert would infer the solution of a problem in the knowledge domain that it is known to the system (Ghani et al., 2009). The inference engine is the component that manipulates the knowledge found in the knowledge base as needed to arrive at a result or solution (Edward and Robert, 1993). According to Amzi (2003), inference engine is the code at the core of the system which derives recommendations from the knowledge base and problem-specific data in working storage. Meanwhile Jocelyn (1996) stated that the inference engine is the generic control mechanism that applies the axiomatic knowledge present in the knowledge base to the task-specific data to arrive at some conclusion. This is the second key component of all expert systems. Having a knowledge base alone is not of much use if there are no facilities for navigating through and manipulating the knowledge to deduce something from it.

The inference engine is the heart of the expert system since this is the part of the program that builds the bridge between information and solutions. Two different approaches to problem solving are usually distinguished and inference engines are accordingly characterized in two different ways, as either backward chaining or forward chaining (Liao et al., 2004). Forward chaining is the process of data gets put into working memory. This triggers rules whose conditions match the new data. These rules then perform their actions. The actions may add new data to memory, thus triggering more rules and so on. This is also called data-directed inference, because inference is triggered by the arrival of new data in working memory. Meanwhile backward chaining is the process when the system needs to know the value of a piece of data. It searches for rules whose conclusions mention this data. Before it can use the rules, it must test their conditions. This may entail discovering the value of more pieces of data, and so on. This is also called goal-directed inference, or hypothesis driven, because inferences are not performed until the system is made to prove a particular goal (Jocelyn, 1996). User Interface

Expert system provides an interface so that user can interact with the expert system. (About Knowledge, 2009). The user interface is the component that allows the user to query the system and receive the results of those queries. User interfaces can be defined as the point where users interact with a computer system (Mockler and Dologite, 1992). The function of user interface is to determine whether the conversation consists of selecting items from menus, responding yes or no to question or filling in forms. The user interface is also responsible for the degree to which the system can explain its solution or otherwise assist users (Zukki et al., 2010).

A machine that is designed to model the behavior of a human expert needs to be equipped with facilities that allow close interaction between the machine and its users. A true expert system should have a user interface designed to operate at a level similar to ordinary conversation. The user interface must allow a user to input data relevant to a problem. This facility must not just be used to obtain the initial knowledge about the case. Once the initial knowledge is matched against the knowledge base, the user must be given an opportunity to provide information that may fill any gaps found to exist in the knowledge. The effectiveness of this interface can determine the quality of the system's response (Navin, 1997).


An expert system may be viewed as a computer simulation of a human expert. Expert systems are an emerging technology with many areas for potential applications. Most applications of expert systems will fall into one of the following categories:

  • Interpreting and identifying
  • Predicting
  • Diagnosing
  • Designing
  • Planning
  • Monitoring
  • Debugging and testing
  • Instructing and training
  • Controlling

The best application candidates for expert systems are those dealing with expert heuristics for solving problems. Conventional computer programs are based on factual knowledge, an indisputable strength of computers. Humans, by contrast, solve problems on the basis of a mixture of factual and heuristic knowledge. Heuristic knowledge, composed of intuition, judgment, and logical inferences, is an indisputable strength of humans. Successful expert systems will be those that combine facts and heuristics and thus merge human knowledge with computer power in solving problems. To be effective, an expert system must focus on a particular problem domain (Media Wiley, 2010).




In this research, an expert system which can provide the public about the Rafflesia species in Malaysia was developed. More understanding about the development stage of expert system itself is needed in order to develop a good system. This chapter consist expert system development stage of ES-RaffID.


Knowledge engineering is an engineering discipline that involves integrating knowledge into computer systems in order to solve complex problems normally requiring a high level of human expertise (Wikipedia, 2010). According to Alison (1994) the knowledge engineer is the AI language and representation expert. Knowledge engineering is a field within artificial intelligence that develops knowledge-based systems. Such systems are computer programs that contain large amounts of knowledge, rules and reasoning mechanisms to provide solutions to real-world problems (Milton, 2003).

According to Dickson (1996), there are two main views to knowledge engineering; the traditional view is known as "Transfer View" and the alternative view is known as "Modeling View". In the transfer view, the assumption is to apply conventional knowledge engineering techniques to transfer human knowledge into artificial intelligent systems. While in modeling view, the knowledge engineer attempts to model the knowledge and problem solving techniques of the domain expert into the artificial intelligent system.

To extract knowledge from the expert the knowledge engineer must first become at least somewhat familiar with the problem domain, maybe by reading introductory texts or talking to the expert. After this, more systematic interviewing of the expert begins (Alison, 1994).


To develop an expert system, we need to follow a few steps starting from task analysis, knowledge acquisition process, prototype development, expansion and refinement and lastly verification and validation.

3.2.1 Task analysis

Identification is the requirements analysis step carried out in traditional software development. It involves a formal task analysis to determine the external requirements, form of the input and output, setting where the program will be used and determines the user. The participants, the problems, the objectives, the resources, the costs and the time frame need to be clearly identified at this stage.

The main objective of task analysis is for the knowledge engineer to identify thus understand the problems to be solved. It is very important to have knowledge that related to the domain because when an organization needs more knowledge on its related field, it will be more crucial for the organization to update and keep in touch with the knowledge (Hendricks and Vriens, 1999).

3.2.2 Knowledge acquisition process

Knowledge acquisition is a process of acquiring the knowledge from human experts or other sources such as books, manuals etc. It involves developing knowledge to solve the problem and transforming it into computer program. It makes this stage is the most difficult to manage and also time consuming. Expert system relies heavily on the knowledge and quality of the knowledge base depends on the level of success achieved in the knowledge acquisition process. Thus, knowledge engineers are responsible to specify all the knowledge needed in order to make the system can function correctly.

Knowledge acquisition is aims to transfer transformation of problem solving and decision making expertise from information source into form for developing knowledge based system (Ghani et al,. 1999). In this process, the knowledge engineer has to studies books, journals, documents, database, and manuals to gather as much knowledge as possible. The more gathered knowledge been analyze is important to make sure the accuracy of the data so that the system can work as professional as an expert in giving solution and advices (Kiong, 2005).

2.2.3 Prototype development

The prototype development is representing how the knowledge base is organized into computer. Usually in development tool process, some criteria must be considered. In choosing an appropriate tool, the aspects that must be considered are generality, testing, accessibility and development tool features. For every aspect, the tools that will be use are different. For example, for testing aspect, by building a very small prototype system, the tool can be tested early.

3.2.4 Expansion and Refinement

In expansion and refinement, the knowledge engineer will expand and refine core knowledge. The prototype will be corrected and expand the knowledge basis of the expert's advice and comments. More information or knowledge is gathered from interviews, proceeding journals, research may be added during this stage. This step is continuously done until maximum satisfactory of the system are achieved. This step also time consuming. After all core knowledge has been gathered and fully expanded, the next step will be testing the system or verification and validation.

3.2.5 Verification and validation

The objective of knowledge verification is to determine the correctness, completeness, and consistency of the system. This stage is divided into two main tasks that are formal tests and test analysis. The test analysis looks for the problems of incorrect answers, incomplete answers and inconsistent answers. It also determines if the problem lies in rules, inference chains, uncertainty, or some combination of these three factors.

The final stage in the development life cycle is knowledge validation which is the most challenging part of building an expert system. The basic motivation behind testing is to control performance, efficiency, and quality of the knowledge base. The goal is compliance with user expectations and system functioning (Awad, 1996). Validation of a knowledge-base system means to make certain that the advice given by the system will be valid in all of its applications. The purpose of this stage is to summarize what has been learned with recommendations for improvements and corrections. The system verification must also be performed in conjunction with all the system knowledge.


While developing the ES-RaffID, the knowledge engineer has implemented all steps or stages in the expert system development. The steps begin with process of task analysis which requires the knowledge engineer to spot the main knowledge that has to be inserted into the system. The process followed by knowledge acquisition process which requires the knowledge engineer to collect all knowledge from many sources such as journals, books, interviews and previous research. After that is prototype development process which the steps of transforming the knowledge gained into computer program. The following step is expansion and refinement which require the knowledge engineer to add more knowledge into the system to increase the efficiency of the system. The last step is verification and validation which require the knowledge engineer to meet the human expert to validate his or her system. The function of validating and verifying the system is to make sure that the system is applicable and reliable thus can be used by the end user.

3.3.1 Task Analysis of ES-RaffID

Task analysis is the first step in developing an expert system. This step requires the knowledge engineer to identify the core knowledge that will be included in the system. At this stage, knowledge engineer will discuss with human expert about important factor or domain that suitable to be inserted in the system.

Knowledge engineering process begins with the analysis of the domain tasks. Task analysis is a methodology tool that can be used to describe the functions of expert performs in solving the problems and to determine the relation of each task to the overall job. In developing an expert system for Rafflesia species identification, ES-RaffID, several tasks were listed which consist of introduction, Rafflesia characteristics, Rafflesia diversity, species identification, habitat of Rafflesia and threats and conservation. The summary of tasks are shown in Table 3.1

3.3.2 Knowledge Acquisition Process of ES-RaffID

The knowledge acquisition process is the transformation of problem solving expertise from some knowledge sources (human expert, textbooks, journals, guidelines, reports, manuals, etc.) into the knowledge base of the expert system. It has proved to be the most difficult component of the knowledge engineering process. It's become known as the 'knowledge acquisition bottleneck', and expert system projects are more likely to fail at this stage than any other.

3.3.3 Prototype Development Process in ES-RaffID

At this stage, the knowledge engineer turns the knowledge into a working computer program. The first prototype should proceed rapidly because of the reasons for implementing the initial prototype is to check the effectiveness of the design decisions made during the earlier phases of development. ES-RaffID was developed by using Macromedia Dreamweaver 8 as the prototype development tool. Development Tool: Macromedia Dreamweaver 8

Adobe Dreamweaver, formerly known as Macromedia Dreamweaver, is a professional web development tool that has been around since 1997. It offers a flexible workspace that can be customized to fit the user's needs and integrates well with other Adobe products. It has a robust collection of tools, yet is still easy for a beginner web developer to use due to its intuitive interface (Catherine, 2010).

Macromedia Dreamweaver 8 is one of the tools for building websites and applications. It provides a combination of visual layout tools, application development features, and code editing support which enabling knowledge engineer at every skill level to create visually appealing, standards-based sites and applications quickly (Iomax, 2008).

Macromedia Dreamweaver 8 offers three different interfaces; design only, HTML code only, or a split screen combination of design and code. Those who are less familiar with HTML may find building web pages in the design only interface much easier. It looks and behaves like a word processor window. For those who like to get into the HTML code and tweak things down to the final tag, the coding window has an easy-to-read format for programming. Tags are color coded, lines are numbered and the nested tags are automatically indented (Catherine, 2010).

Most tools in Macromedia Dreamweaver 8 are located on the main screen so they are easy to find. The toolbar at the top of the screen can be set for menus or tabs and sorts tools into several categories based on their purpose (Catherine, 2010).

3.3.4 Expansion and Refinement

This step involves evaluating the performance and utility of the prototype program and revising it as necessary. It involves checking for mistakes in knowledge acquisition to ensure that the knowledge base correctly reflects that of the expert, and to establish that the system performs with an acceptable level of accuracy, user-friendliness, and overall usefulness. During develop the ES-RaffID, an editing processes were continuously done. The knowledge engineer kept on searching for more knowledge to be added into the system. The knowledge engineer also checked the system and the prototypes to make sure that no typing error and no mislead of knowledge in the system. This was done in order to ensure that the system can act perfectly as human expert. The process of editing the system terminated after the knowledge engineer satisfied with the system.

3.3.5 Verification and validation

The process of verification and validation were done by the human expert to confirm that the system can function correctly and is applicable to the real world. In this step the knowledge engineer have to find several human experts in order to ensure that the system can act or give the suggestions or recommendations just like human experts do. The verification process was done to make sure that the system is built right. While doing the verification, the human expert will checked whether knowledge base of the ES-RaffID is complete or not and to ensure that the inference engine can properly manipulate the information.

The validation of ES-RaffID was conducted to confirm that the knowledge base in the system is correct and provable. This process also to determine that the ES-RaffiD is complete and can performs the function in very well. It is also to ensure that ES-RaffID usable for the intended purposes.




Expert System for Identification of Rafflesia species in Malaysia or so called ES-RaffID is created for identification of Rafflesia species purpose. This chapter will discuss more detail about the architecture of ES-RaffID and also the flow of developing the system. To ensure the prototype is applicable and also can give accurate recommendation as well as an expert, lots of knowledge has been acquired from many sources such as journals, books, articles and website.

In addition, the discussion on IF-THEN rules also will be included as it is the core knowledge in the ES-RaffID. Other than that the user interface and targeted user will be clarified in this chapter.


The knowledge base of ES-RaffID prototype consists of two main modules:

  • Selection of location the Rafflesia has been found
  • Selection of characteristics based on Rafflesia appearance

Each module has its own purpose. The purpose of module for selecting of Rafflesia location is to focus on Rafflesia species on that particular location due to some Rafflesia species are endemic to certain place only. The function of the second module is to identify the Rafflesia species that match with the characteristic given by users. The system architecture of ER-RaffID can be expressed as shown in Diagram 4.1.

The process of identifying Rafflesia species begins once the users choose the location of the Rafflesia were found and the appearence characteristics of the Rafflesia. After the users completed the questions given, the prototype will then show the specific name and also the criteria of the Rafflesia species based on the characteristics given by the users.


Principally, the ES-RaffID process consists of two different phase of question. The phases are as follow:

  • Location towards Identification
  • Characteristics towards Identification.
  • In the first phase, the selected location will determine a group or population of Rafflesia species according to their habitat. The users only have to choose the location of the Rafflesia were found before the prototype brings them to the next phase of question.
  • The following phase is much related to the first phase. The users have to detail out the appearance characteristics of the Rafflesia. This include the colour of perigone lobes, the number of "windows", the diameter of open flower, the characteristic of blots or warts and also the altitude of distribution. The process of detail out will be done by answering each question given by the prototype. The reason this phase was done because of each Rafflesia species have different appearance characteristics and the species can be differentiate by spot the physical characteristic of it.
  • After the users finish answering all questions, the prototype will lead the user to the result which shows the scientific name of the Rafflesia and also additional information on that particular Rafflesia species. The flow of ES-RaffID processes is shown in Diagram 4.2.

    4.4 MODULES IN ES-RaffID

    There are five modules in ES-RaffID; Introduction, Rafflesia Characteristic, Rafflesia Diversity, Rafflesia Habitats, Treat and Conservation. The Introduction module is the main interface that will appear after the users start conversation with ES-RaffID.

    4.4.1 Introduction module

    Figure 4.4 shows the first interface that users see when they enter ES-RaffID. There are seven modules in the main interface; Homepage, Introduction, Rafflesia Characteristic, Rafflesia Diversity, Rafflesia habitats, Threat and Conservation and also the Species Identification. When the user clicks over one of the button such as Rafflesia Characteristic, the prototype will run to the interface that has been connected with the button.

    The Introduction interface is about general knowledge of Rafflesia. Rafflesia's family name is Rafflesiacea which consist of eight genuses in this family. The genus Rafflesia is named after the adventurer and founder of the British colony of Singapore, Sir Stamford Raffles. The best known of Rafflesia species is Rafflesia arnoldii also has been described in this interface.

    4.4.2 Rafflesia Characteristics Module

    Figure 4.5 shows the Rafflesia characteristics interface which consists of all characteristic of Rafflesia. This interface also connected to the Rafflesia morphology as well as the host plants. Rafflesia comprises of several characteristics that differentiate it from other plants. The petal of Rafflesia or so called perigone lobes can reach up to 1 meter in diameter. The flower gives off rotten flesh smell which attract the carrion fly which then the pollination process begin. The Rafflesia is a parasitic plant and its host plants are vines of Tetrastigma species.

    4.4.3 Diversity of Rafflesia Module

    Figure 4.6 shows the diversity of Rafflesia interface which consist of known and unknown species of Rafflesia and their locations as recognized by Meijer in 1997. According to Meijer (2007), there are 23 known species of Rafflesia around the world. Some of the Rafflesia species are endemic to certain location. For example, Rafflesia tengku-adlinii only can be found in Sabah while Rafflesia azlanii only can be found in Peninsular Malaysia.

    4.4.4 Habitat of Rafflesia Module

    Figure 4.7 shows the habitat of Rafflesia interface. Map of Rafflesia species location is connected to this interface. The Rafflesia can be found at altitudes of between 150 and 1450 meters in the forests of Malaysia, Southern Thailand, Sumatra and Java in Indonesia. In these tropical rainforests, the climate is continuously warm and humid, with very high humidity frequently reaching 100% at night. The Rafflesia is rare and fairly hard to locate.

    4.4.5 Threats and Conservation of Rafflesia Module

    Figure 4.8 shows the threats and conservation of Rafflesia interface. It consists of Rafflesia threats and ways to conserve it using in-situ conservation and as ex-situ conservation. The main threats of Rafflsia are the indeginous activities that collect Rafflesia bud for medication purpose. The Rafflesia buds are sold to mothers after giving birth and used to recover faster. The other threat comes from visitors that eager to see the Rafflesia in bloom. One of the factors the Rafflsia becomes endangered species is because the visitors or tourist cause massive trampling, even to level where some populations are trampled to extinction. There are several methods to conserve Rafflesia such as by using in-situ approach and ex-situ approach.


    The ES-RaffID is applying IF-THEN rules in developing the prototype. The IF-THEN rule of the system can be simplified as follows:

    Figure 4.9 shows the Rafflesia species identification interface. The ES-RaffID only covered identification of Rafflesia in Malaysia. There are eight out of 23 species of Rafflesia is located in Malaysia. Once the users click "Run" button, another interface will appear.

    Figure 4.10 shows the first interface that will appear after the user run the system. The prototype will ask the user about location the Rafflesia was found. Each alternatives answer chosen by users will lead to the next interface. For example, if the users click the button "Borneo (Sabah/Sarawak)", the prototype will lead to the interface that connected to it. After choose the location of Rafflesia was found, the prototype will lead to the interface which ask the users about the characteristic of the Rafflesia.

    Figure 4.11 shows the interface which the users need to answer the appearance characteristics of the Rafflesia. After answer this question, another interface will appear which asking the users relatively same kind of question that is about appearance characteristics of Rafflesia. The users need to answer all questions given by the ES-RaffID to find out the specific name of the Rafflesia species.

    Figure 4.12 shows an example result of Rafflesia species interface. After the users answer all questions given, the system will shows the answers that the users gave and stated the specific name of the Rafflesia. The additional information about the Rafflesia species are also included in the result interface after the users click on the name of Rafflesia species. For example of interface in Figure 4.12, the users have to click on the "Rafflesia pricei" to find out more detail about this species.

    4.6 End Users

    After the system was completely developed and verified by human expert, it is then readily to be used by the end user. The end users of ES-RaffID are as follows:

    • General public
    • Naturalist
    • Conservationist of endangered species

    The ES-RaffID is useable by general public as a method to expose the public about the Rafflesia species in Malaysia. It is also can be a useable guide or system for conservationist of endangered species and naturalist to identify the Rafflesia species.




    Expert system is a vast field to be explored especially in environmental field. This system will help the user to make decision and also give recommendation to the users about the action to be taken. In this research, the expert system technology is purposely for identification of Rafflesia species in Malaysia. The expert system technology can be category as a new technology in Malaysia and it is very appropriate methods in conservation of species processes.


    The ES-RaffID was developed by using Macromedia Dreamweaver 8. There are a few advantages using Macromedia Dreamweaver 8 as a prototype tool. The following are examples of the advantages using Macromedia Dreamweaver 8.

    • Macromedia Dreamweaver 8 is very easy to be used by a new or inexperienced computer programmer.
    • Macromedia Dreamwever 8 also has simple toolbar; menu toolbar, insert toolbar, document toolbar and standard toolbar. Menu toolbar contains standard menu items for File, Edit, view, Modify etc. Insert toolbar is used to insert various objects such as images, tables, links, horizontal rules and line breaks. Document toolbar displays buttons frequently used such as View, Title field, Refresh and Preview in Browser options. While Standard Toolbar has file management shortcuts such as New, Open, Save, Cut, Copy, Paste, Undo and Redo.


    Despite the research done on expert system, there are still limitations in the system that needs improvement in the upcoming research. A few recommendations for an ES-RaffID to make the system applicable in tha future are as follows:

    • Firstly, the system must be update from time to time following the latest found or research about the Rafflesia species. More information about the Rafflesia species and the pollination process can be added.
    • Secondly, more knowledge base about Rafflesia species in other South-east Asia can be added to the system. This makes the system more complete and can be used to identify all Rafflesia species.
    • The performance of user interface could be enhanced by system modification in few aspects such as more attractive design and colours. Besides combination of suitable colour of background and interface will make the system more interesting.


    • About Knowledge. 2009. Components of Expert System.
    • Alison. 1994. Expert System Architecture.
    • Amzi. 2000. Introduction to Expert System
    • Attenborough, D. 1995. The Private Life of Plants. BBC Books, London.
    • Awad, E. 1996.Building expert systems: principles, procedures, and applications, West Publishing Company pp. 372-398.
    • Beaman, R., P.J. Decker and J.H. Beaman. 1998. Pollination of Rafflesia. American Journal of Botany 75(8): 1148-1162
    • Carol, E. B. and E. O. Daniel, 1995. Introduction to Artificial Intelligence and Expert System.
    • Catherine, C. 2010. About Dreamweaver Software.
    • Cohen and Feigenbaum. 1982. P.R. Cohen and E.A. Feigenbaum, Editors, The Handbook of Artificial Intelligence Vols. I-III, Kaufmann, Los Altos, CA (1982), p. 1476.
    • David, L. W. 1996. Bringing Common Sense, Expert Knowledge, and Superhuman Reasoning to Computers
    • Dickson, L. 1996. Knowledge Engineering Part A: Knowledge Representation. Department of Mathematics, Statistics, and Computer Science. UNE, Armidale.
    • Earlham College. 2003. The Genus Rafflesia
    • Edward, F. and S. E. Robert. 1993. Chapter 1: Introduction of Expert System and Artificial Intelligence
    • Generation5. 2009. Expert System
    • Ghani, A., E. A. Aljack, M. Daud and M. Z. Bardaie. 1999. Developmet of Expert Database System for Environmental Impact Assessment (EDEIA). Malaysia Journal of Computer Science 12(1):19-27
    • Ghani, P.H.A., L. Abdul Manaf, M. Kamil, & Mohamed. 2009. Knowledge-Based System for River Water Quality Management. European Journal of Scientific Research. Universiti Putra Malaysia, Malaysia
    • Iomx. 2008. Macromedia Dreamweaver
    • Jocelyn, P. 1996. How the Inference Engine Works.
    • Kashyap. N. 1997. An Expert Systems Application in Electromagnetic Compatibility. Master's thesis, University of Missouri, Rolla.
    • Kiong. S. W. 2005. MobES Expert System. (cited on 30 March 2008).
    • Liao, H. T., D. Enke and H. Wiebe. 2004. An expert advisory system for the ISO 9001 quality system. Expert System with Applications 2004; 27: 313-22.
    • Media Wiley. 2010. Fundamental of Expert System
    • Milton, N. 2003. Knowledge Engineering
    • McEneaney, J.E., 1992. Computer-assisted diagnosis reading: an expert system approach.Journal of Reading, pp. 36-45.
    • Meijer, W. 1997. Rafflesiaceae. In: Flora Malesiana, Series I, Vol. 13, 1-42
    • Mockler J, and G. Dologite. 1992. Knowledge-based system: an introduction to expert systems. Mcmillan, New York
    • Mohd Zukki, N. I., L. Abdul Manaf and A. S. Mohd Armi. 2010. The Development of an Expert System for Decision Making in Forest Resources Managemant. UPM Serdang, Selangor, Malaysia
    • Nais, J. 2001. Rafflesia of the World. Sabah Parks, Kota Kinabalu, Malaysia
    • Tanimoto. 1987. S.L. Tanimoto, The Elements of Artificial Intelligence. In: Computer Science Press, Rockville, MD (1987), p. 527.
    • Taylor, J. 2008. Dreamweaver 8 New Features and Benefits. Turban and Aronson. 2001. PC Electronic Cards Diagnoses Using Expert Systems
    • Turban, E., Rainer Jr., R.K. and Potter, R.E. 2001.Intelligent systems in business, introduction to information technology, Wiley, New York pp. 352-386
    • UNEP-WCMC: Kinabalu Park (cited on February, 2008).
    • Walter, K. S. and H. J. Gillett. 1998. 1997 IUCN Red List of Threatened Plants.IUCN, Gland, Switzerland and Cambridge, UK.
    • Wentworth, J.A. 1993. Highway applications of expert systems.Public Roads564, pp. 148-152.
    • Western Michigan University, 2003. Rafflesia.
    • Wikibooks. 2010. Components of Expert Systems.
    • Wikibooks. 2010. Introduction to Expert System. (2010) Wikipedia. 2010. Expert System.
    • Wikipedia. 2010. Knowledge Engineering.
    • Wikipedia. 2010. Knowledge Representation and Seasoning 2010. (2010)
    • Zuhud, E. A. M., A. Hikmat & N. Jamil. 1998. Rafflesia Indonesia, Keanekaragaman, ekologi dan Pelestariannya. Yayasan Pembinaan Suaka Alam dan Suaka Margasatwa Indones (The Indonesian Wildlife Fund) dan Laboratorium Konservasi Tumbuhan, Institu Pertanian Bogor, Indonesia.