The Expert System of Determining the Type of Malaria by using Dempster-Shafer Method

-Malaria is the most dominant disease in Asia and Africa and may become a life-threatening disease for it suffers. The types of malaria such as Plasmodium Vivax, Plasmodium Ovale, Plasmodium Malariae, and Plasmodium Falciparum are mostly infected people around the world. These types of malaria have certain symptoms that drives difficulties for some patients to confirm which malaria that their infected. A clinical testing and medical diagnostic assessments may be performed to determine the types of malaria, but utilizing a system also brings some benefits for rural areas which lack of medical facilities. The study develops a system by implementing the Dempster Shafer method to determine types of malaria. We collected the knowledge from the experts including 18 possible symptoms along with the density value. This paper present 5 cases of sufferers and provide the system result with the possibilities of malaria types. The result pointed out a various percentage of malaria types that may infected to the patients.


INTRODUCTION
Malaria is kind of disease that has been spread all over the world, especially on the Africa and Asia.Malaria is also a disease that is quite dangerous and can result in death for its sufferers [1].Based on the report from the World Malaria Report by WHO, the number of sufferers of this disease has reached 247 million people in 2021 with a total death of 619,000 fatalities.Hence, it was indicate an increase in the number of sufferers compared to the number of sufferers in 2020 which has reached 245 million sufferers and 625,000 deaths.[2].Malaroa is an infectious disease that is usually caused by parasites in the form of sporozoites of the genus Plasmodium and is usually transmitted to humans through the bite of an infected female Anopheles mosquito with the parasite.[3].The disease transmission occurs when an infected mosquito bites a person, then injects the malaria parasite into their bloodstream Then, the parasite attacks and destroys red blood cells causing the characteristic symptoms of malaria.This mode of transmission is a significant public health concern, particularly in areas where malaria is endemic [4].the symptoms commonly experienced by sufferers of this disease are fever, chills, dizziness, and nausea.Furthermore, if it is not treated immediately it will be fatal which can cause severe anemia, kidney failure and even death [5].There are various types of malaria spread throughout the world, including Plasmodium Vivax, Plasmodium Ovale, Plasmodium Malariae, Plasmodium Falciparum.These types of malaria have different symptoms and so does the impact of the disease [6], [7].Plasmodium Vivax, Plasmodium Ovale, and Plasmodium Malariae are types of malaria that are quite harmless for human health.In other words, it has a small possibility of having an impact on death, but proper management is necessary to avoiding the chronic malaria [8].On the other hand, Plasmodium Falciparum is the most dangerous and life-threatening type of malaria which results in the death of most sufferers.[9].Moreover, the Plasmodium vivax and Plasmodium falciparum are types of malaria that is most frequently found in Indonesia [10].These various types of malaria drive a proper method to determine which type of malaria a person suffers from.Ensuring the patient receives the most effective medication and care is of utmost importance, as it significantly enhances the likelihood of a successful recovery.Additionally, the precise identification of the malaria strain plays a pivotal role in preventing the spread of the disease.Determining the type of malaria and evaluating its impact are usually conducted clinical testing and medical diagnostic assessments.However, there are such mathematical calculations that are useful for understanding and managing malaria which can later be applied to systems, such as expert systems.This mathematical calculation can be used for various purposes, including inference, decision making, handling, uncertainty, and knowledge representation [11].As one of mathematical analysis, the Dempster-Shafer (DS) method can be used for reasoning under uncertainty and combining evidence from multiple sources.This method usually provides a way to model and measure the level of confidence or uncertainty in a systematic way [12].Dempster's combination rule is used to combine the belief function, namely by calculating a new belief function by considering the intersection and overlap of the belief function related to various sources of evidence.This resulting combined belief function provides a more comprehensive representation of the overall beliefs or uncertainties about the probable outcomes [13].The DS method has been widely used in various fields, including artificial intelligence, decision support systems, and expert systems.This method can assist in solving problems related to information that is uncertain or conflicting.Thus, enabling better decision making and inference under uncertainty [14].The DS method is usually also widely applied to expert systems related to the diagnosis of human diseases [15]- [17] as well as plants diseases [18]- [20], and many other fields.The DS method is widely recognized as the most commonly employed approach in expert systems across diverse domains.The present investigation endeavors to ascertain the specific type of malaria afflicting an individual through the utilization of the DS method.The types of malaria that used are Plasmodium Vivax, Plasmodium Ovale, Plasmodium Malariae, and Plasmodium Falciparum, with reference to the symptoms experienced.The research conducted also involved the creation of an expert system that effectively incorporated the DS methodology.

RESEARCH METHODOLOGY 2.1 Research Steps
A structured research phase is set to ensure the research is run properly.The implementation of the research will be conducted in a sequential manner, following each of the outlined stages.These stages have been visually represented in Figure 1 for clarity and ease of reference.The study will adhere to a structured approach to ensure the accuracy and reliability of the results.The sequential implementation of the research stages will enable a comprehensive analysis of the data collected, leading to a thorough understanding of the research topic.

Figure 1. Research Steps
Based on Figure 1, there are several stages involved in carrying out this research.The first stage is identifying the problem, which is the process of defining the problem to be investigated which is useful for establishing fundamental aspects of research [21].In this study, the problem identified was how to determine the type of malaria suffered by some patients using the DS method and implemented it on an expert system.The second stage is data collection, namely the stage of collecting relevant data to answer research objectives from various sources [22].Subsequently, the third stage of the project will entail the integration of the DS method into an expert system, which will be followed by a comprehensive evaluation of the outcomes achieved through the application of the method.

Data Collection
In this study, the data collection step is performed using the interview method, which is explores and understands the perspectives and beliefs of the informants in depth.This method is used to investigate more detailed responses and clarify ambiguous or unclear points [22].The interviews were conducted with doctors as resource persons, then the results of the interviews will be used as input in the form of knowledge.The data obtained from informants included the types of malaria, its symptoms and their beliefs, and the rule base to describe the type of malaria.Table 1 portrays the types of malaria studied along with the symptoms.Furthermore, Table 2 explains the density value of each possible symptom experienced by malaria sufferers.

Dempster Shafer Method
Dempster Shafer (DS) is a method that carries a measure of the strength of evidence in supporting a set of propositions called as belief (bel).This belief function lies in a value of 1 or 0, which is a value of 0 indicates that there is no confidence in the evidence, and 1 indicates certainty in the evidence [23].The DS method uses plausibility (Pls) as a parameter of the uncertainty value over an evidence which will affect the level of confidence of the evidence.For example, if there is a belief in something (X), then the value of bel(X) is 1, so the value of Pls(X) will be 0. Equations ( 1) and ( 2) below are used to calculate the value of bel and Pls [15].
The DS method also carries a universal set of hypotheses that are summarized in an environment called a frame of discernment (θ), in this θ there are elements in the form of possible answers and there is also the possibility of one required answer or what is commonly called the probability function density/mass function (m).If X and Y are subsets of θ, with m1 as the density function of X and m2 as the density function Y, then m3 can be formed as a combined function of m1 and m2 which is formulated as in Equation (3).

System Implementation
The expert system that was built has several menus including Login Page, Home Page, Diseases Page, Symptoms Page, Diagnostics Page, Information Page and others.The menu displays that have been built are separated between admin and user view.The Login Page is used by admin to enter the system by inserting certain username and password that will be filled in by the system's admin.Figure 2 illustrates the system's Login Page interface.Moreover, The Home Page is used to display all menus in the system.Users are able to select the pages to jump to another page.The Home Page is depicted in Figure 3.  Furthermore, the system also presents the Disease Page, which is used to process disease information in the system.The information processed is the name of the disease, disease details and disease advice.Figure 4 shows the interface of Disease Page from the system.

Figure 4. Disease Page Interface
Then, there is a Knowledge Page that used in the system as a rule base for malaria.The page is quite important due to its function to determines the diagnosis result.The data processed by Admin in this menu is the value of trust and distrust given by experts.Knowledge Page interface is displayed in Figure 5.Moreover, the system also has the Diagnosis Page that used to determine the type of disease according to the symptoms experienced by the user.In this page, the user will select the symptoms and select the conditions experienced, after selecting the symptoms and conditions the user must select the search button.Then, the system will display the next menu which is the Diagnostic Results Page.

System Result Evaluation
The system is used to determine patients type of malaria according to the symptoms that listed in Table 3.Firstly, we input all the symptoms into the system of each case.The system will do the calculation by using DS method including Equation ( 1), (2), and (3) that have been implemented on it.Then, we collect the result based on the system output, which is clarifying which types of malaria that suffered by the patient.Table 4 shows the result of system output.

CONCLUSION
Malaria is one of harmful disease which spread all over the world.Malaria is become a pernicious disease that has a global reach.It manifests in a multitude of symptoms and requires a range of treatments depending on its various types.The disease is caused by the Plasmodium parasite, which is transmitted through the bite of infected Anopheles mosquitoes.The World Health Organization has implemented various strategies to combat the disease, including the distribution of insecticide-treated bed nets and the use of antimalarial drugs.This disease has numerous symptoms and various handling according to it types.Hence, a system is necessary to help patient getting information regarding types of malaria that may infected to them.This research endeavors to create an expert system that can accurately identify the various types of malaria that may afflict patients.The system is designed to utilize Dempster Shafer method to analyze patient data and provide accurate diagnoses.By leveraging the latest technological advancements in the field of medical diagnosis, this study aims to improve the accuracy and efficiency of malaria diagnosis, ultimately leading to better patient outcomes.The expert system developed in this research has the potential to revolutionize the way malaria is diagnosed and treated, providing a valuable tool for healthcare.The system allows users to input the symptoms that their experienced into several forms, then the system presents the malaria type that may possibly occur.The Dempster Shafer is used as the method of system and the study also tested the system with 5 cases of patients and resulting various outcome.The research expect that other methods could be implemented to compare or optimizing the result, such as Fuzzy Logic or Genetic Algorithm.

Figure 2 .
Figure 2. Login Page Interface

Figure 3 .
Figure 3. Home Page Interface Figure 4. Disease Page InterfaceThen, there is a Knowledge Page that used in the system as a rule base for malaria.The page is quite important due to its function to determines the diagnosis result.The data processed by Admin in this menu is the value of trust and distrust given by experts.Knowledge Page interface is displayed in Figure5.Moreover, the system also has the Diagnosis Page that used to determine the type of disease according to the symptoms experienced by the user.In this page, the user will select the symptoms and select the conditions experienced, after selecting the symptoms and conditions the user must select the search button.Then, the system will display the next menu which is the Diagnostic Results Page.Figure6portrays the interface of Diagnosis Page of the system.

Figure 7 .
Figure 7. Diagnosis Results Page Interface

Table 1 .
Types of Malaria and the Symptoms

Table 2 .
Symptoms dan Density Value

Table 3 .
Cases and the symptoms

Table 3 .
Cases and the symptoms

Table 4 ,
there are various percentages of malaria are produced by the system from the cases.Patient number 1 and 5 are infected with the Plasmodium Tertiana with the possibility percentage of 73% and 91%.Moreover, patient number 2 is suffered of Plasmodium Ovale with the possibility percentage about 81%, then patient number 3 is may exposed to Plasmodium Tropika with the possibility percentage of 91%.Last but least, patient number 4 has low possibility of Plasmodium Malariae with the percentage of 56%.