Computational Intelligence Method for Early Diagnosis Dengue Haemorrhagic Fever Using Fuzzy on Mobile Device

. Mortality from Dengue Haemorrhagic Fever (DHF) is still increasing in Indonesia particularly in Jakarta. Diagnosis of the dengue shall be made as early as possible so that first aid can be given in expectation of decreasing death risk. The Study will be conducted by developing expert system based on Computational Intelligence Method. On the first year, study will use the Fuzzy Inference System (FIS) Method to diagnose Dengue Haemorrhagic Fever particularly in Mobile Device consist of smart phone. Expert system application which particularly using fuzzy system can be applied in mobile device and it is useful to make early diagnosis of Dengue Haemorrhagic Fever that produce outcome faster than laboratory test. The evaluation of this application is conducted by performing accuracy test before and after validation using data of patient who has the Dengue Haemorrhagic Fever. This expert system application is easy, convenient, and practical to use, also capable of making the early diagnosis of Dengue Haemorraghic to avoid mortality in the first stage.


Introduction
Health is a priceless treasure, so we are willing to do everything to maintain physical health.Anyone may face health problems, regardless of age, sex, and profession.
Deaths due to dengue hemorrhagic fever (DHF) in Jakarta are still high.Active surveillance of hospital reports Jakarta Health Agency of 2004, a 20 643 dengue cases with 91 deaths soul or case fatality rate (CFR) of 0.44%.Dengue deaths in Jakarta are quite high compared dengue deaths were reported in the world Organization World Health (WHO) 2000 CFR of 0.16%.Deaths due to many factors, among others, due to late diagnosis' [4].
Dengue Hemorrhagic Fever (DHF) is one of health problems that affect productivity of each person in fact possibly end in death.Dengue Haemorrhagic Fever symptoms have been identified by common people based on the general known physical characteristic without supported by other medical fact or medical consideration.(Unfortunately, some other disease has the symptoms that sometimes resemble those of Dengue Haemorrahagic Fever so that the patient possibly is misdiagnosed).Consequently, the patient is treated in the wrong way and do not recover.The laboratory test can be performed to get a better diagnosis but this way is relative expensive and take longer time to know the results, besides the adequate laboratory does not exist in every area in Indonesia.
A best way is to consult an expert or specialist doctor.Although at the present there are many experts or specialist doctor, the obstacles still there impede patient to get proper treatment, that are the limited visiting hour and the medical treatment centre is not easy or near to reach so that many patient waiting in queue whereas they have to be treated promptly.In this situation, the common people as users need an expert who can facilitate them to get early diagnosis to get promptly treatment in first stage.One way to solve the problem is make a tool that can take a role as an expert who is capable of making accurate diagnose and that can be run on mobile platforms.
Mobile technology development increase rapidly, both software and hardware.This condition facilitates the public, especially mobile or smart phone users to obtain any information at any time and any place with fast and practical way [1].Expert system is one branch of artificial intelligence that studies how to adopt the way an expert thought and their logic in solving a problem, and make a decision or conclusion based on a number of facts.The basis of expert systems is how to transfer the knowledge possessed by an expert into a computer, and how to make decisions or draw conclusions based on that knowledge [5].Method of expert systems such as the fuzzy logic system has advantages than other methods of expert system such as artificial neural networks.Fuzzy logic system can tolerate inaccurate data, able to model a very complex nonlinear function and can establish and apply the experiences of experts directly without having to go through the training process.

Methodologies
The following is scope of the study: x Designing, manufacturing, and testing expert system applications that can help diagnose dengue fever through Android-based mobile devices with Sugeno fuzzy inference method based on four main symptoms, namely: fever, skin rush, spontaneous haemorraghing and tourniquet test.
x Design and manufacture of web-based administration system, as a means to exercise control and supervision of mobile-based applications.

Design
Users run applications that have a diagnosis of dengue fever in the mobile phone Android operating system.
Then the user has to answer some questions provided by the application to obtain data that will be processed.Once collected answers are entered then the application will proceed to diagnose whether the user is suffering from dengue fever or not.Users can also update the application in order to continue receiving the latest knowledge for further development.Users also perform data transmission diagnosis.

Input
The input is the answers given by the application user to the questions.Further answers will be treated to manage all of its value in accordance with the membership function so that the value becomes a fuzzy value.There are three membership functions contained in the application of each criterion.Then after getting the value of each fuzzy membership function of each category, the value is incorporated into the rules (rules) that exist [2].

Fever
There are three parameter of fever namely low, average, and high.The fever level is evaluated based on human body temperatur [5].The temperature measurement is between 36°C-42°C.The body is considered suffering from fever if its temperature is 37°C or higher.The fever fuzzy score based on body temperature measurement shall be as follow:

Skin Rash
There are three parameter of skin rash namely a few, moderate, and many.Evaluation of the skin rash is determined based on the amount of petekia (small skin rash) per circle with diameter 2.8 cm on the arm, chest or face [5].The petekia measurement result will be scored between 0.00 -1.00.The level of skin rash is determined based on score as follow: Table 2.The Skin Rash Fuzzy Score

Spontaneous Haemorraghing
There Are three parameter of spontaneous haemorraghing namely indistinct, distinct, and very distinct.The evaluation of spontaneous haemorraghing level based on a haemorrhage occurred inside nose, gum, melena, and hematemesis [4].The result of spontaneous haemorrhage will be score 0.00 -1.00.The level of spontaneous haemorrhage is determined based on score as follow:

Torniquet Test
There are three parameter of torniquet namely negative, indicisive, and positive.The evaluation of torniquet test level is determined by amount of petikia that observed per circle with diameter of on fossa cubiti [5].The counting result of petekia will be scored between 0.00 -1.00.The torniquet level will be determined based on score as follow :

Defuzzification Method
This research used the weighted average method as Defuzzification method.In weighted average method, each membership function is formed multiplied by the result of a rule that exists and everything added up, then the value is divided by the sum of all the membership function [3].Sugeno defuzzification Weight Average (WA) : Whereas: α = fire strength

Output
The result of diagnose is membership function from range 90 up to 91.The score included in this range will be evaluated and lead to the resume opstions namely infected by DHF or not infected.Range of score will be classified into three categories: DOI: 10.1051/ C Owned by the authors, published by EDP Sciences,

Table 4 .Fig. 8 .
Fig.8.Membership Function Uji Torniquet Tipe Gaussian possibility that you are infected by DHF.You need to visit your doctor to have medical examination.More than 90.4 up to less than 90.6.There is high possibility that you are infected by DHF.90.6 -91.09ImplementationThis program can be run with minimal configuration of mobile devices:

Table 1 .
The Fever Fuzzy Score

Table 3 .
The Haemorrhage Fuzzy Score