Expert System for Diagnosing Newborn Babies Disease Using the Sorgenfrei Similarity Method

ABSTRACT


I. INTRODUCTION
Neonatal is defined as a clinical syndrome in aninfant 28 days of life or younger [1].Newborn children screening is a must for examining every newborn for persuaded dangerous or potentially deadly disorders that aren't otherwise obvious at birth [2]. Weightless or below average becomes a problem that often occurs in newborns. In an emergency, diagnosis and help are needed quickly. Delay in providing help can result in death to the baby. In 2012 the number of neonatal infant deaths was 19 deaths per 1000 live births. In 2017 there were 15 deaths per 1000 live births [3]. In 5 years there was only a significant decrease in mortality.
Lack of adequate health facilities, inadequate diagnosis, and still minimal human resources in the field of health, causes a high rate of newborn mortality. These events, more frequently caused by improper intrauterine transfer of maternal blood oxygen to the fetus resulting in an acute inadequate supply of oxygen to the fetal organs and tissues [4]. Data collection starting from handling pregnancy, managing data of health workers, managing information on pregnancy problems by the community, delivery by medical personnel and neonatal visits need to be managed properly to produce information about good handling so that neonatal recovery can be adjusted by taking action -appropriate measures for neonatal issuance based on information data that has been generated with these problems, the authors want to create an expert system that is useful for diagnosing newborn diseases.
In the case of report submission, there is a problem caused by the delay in submission of the report which causes the data obtained is not up to date, so the decision making based on the data becomes inaccurate.To provide information about the disease accurately and precisely needed a system that can do this simultaneously. The most appropriate system used for this problem is the expert system.
Expert systems are computer programs that are derived from a branch of computer science research called Artificial Intelligence (AI) [5]. Expert systems are used to solve a large number of problems such as decision-making [6]. Computer-based systems using knowledge, facts and reasoning techniques for an expert to solve problems. The inference engine applies the rules to known facts to deduce new facts [7].From the problem the current proposed expert system is specialized in the diagnosis ofdisease in newborns [8].

A. Similarity Algorithm
A similarity algorithm is a way to calculate the similarity of two objects by comparing the similarity based on syntax or semantic values. The coefficient is adopted to display the quantity of similarity. Algorithms for solving many pattern recognition problems such as classification, clustering, and data retrieval problems for more details can be seen in Table1.

B. Similaritas Sorgenfrei
The similarity is a measure of the closeness of an object with other objects. Then what is meant by Sorgenfrei similarity is the method used to calculate the similarity between the distance of two probability distributions with the formula referred to in equation (1).
Information: S = Value Similarity a = The same number of attributes between consultation and data stored in the database. b = The amount of data stored in the database, which is not in the consultation. c = Number of consulting data attributes, which are not stored in the database.

C. System Description
Website-based expert system for newborns using the Sorgenfrei method. This application allows users to view newborn disease information. To start the application process, it is divided into two access rights, namely user and admin. The user can consult with the selected symptoms experienced by the baby. While the admin can make a consultation like a user plus the admin can input information, symptoms, and diseases.
In implementing this child expert system, it takes 4 stages of the process, namely retrieve, reuse, revise, and retain. The retive process is the process of finding data in a database with method sorgrenfrei. Then the reus system process that provides the results of the calculation of the greatest similarity value will be used as a solution to the disease. Furthermore, in the process of revisingprocess of revising the values that come out of the proposed solutionreus process that still does not meet the requirements will be accompanied in a special table, for later repaired experts to find the right solution [9]. After the revision process is finished and the most appropriate solution is found, the expert will add the relation to the relation table which can later be used for solving similar cases in the future [10].
Admin's work process begins with logging in to the login page, then the admin will be directed to a display that has a choice of menus, illnesses, symptoms, solutions, and relationships to maintain data and information. For more clearly how the system works User work system architecture can be seen in Picture 1, while the admin work systemarchitecture can be seen in Figure 2 .

97
The picture above explains the flow of the user's role in using the system to consult newborn diseases In the figure 2 above is an admin role that can access all the menus or features that are in the system created.

D. Knowledge Representation
The symptoms table is used as a knowledge base that will be compared with many symptoms, both those that have been recorded and those that have never been recorded. This database is used as a basis for comparing a user's consultation with a pre-existing knowledge base. In the symptom table, there is a symptom code that shows the sequential number of symptoms and the name of the disease symptoms that exist in newborns. For more details, can be seen in Table 2. Clinical syndromes G23 Active movement (looking healthy) G24 Self limited disease G25 Responsifterhadapterapisuportif yang diberikan (kesadaranbaik) G26 Cough G27 Sometimes sneezing G28 Secretions from the nose (runny nose) G30 Body weight at birth LBW = 1500 -2499 grams G31 Body weight at birth BBLSR = 1000 -1499 grams G32 Weight at birth BBLER = <1000 grams G33 Body length <45 cm G34 Head circumference <32 cm G35 Chest circumference <30 cm G36 Inactive movements (hypotonic muscles) G37 Head bigger than body, thin and fine hair G38 Soft skull bones G39 Simple-shaped ears and little cartilage G40 small / not formed nipples G41 breathing <20x / minute G42 Thin, transparent skin, lanugo (fine hair), especially on the forehead, forehead temples and arms, visible blood vessels G43 Genetalia is not perfect, in women the labia minor has not been covered by labia majora, the clitoris protrudes. In small scrotal LK men, the testes are not palpable G44 Reflexes suck, swallow and cough weak G45 Edema Extremities (swelling) G46 Convulsions G47 Fever (temperature> 37.5ºC) G48 Irregular eye movements G49 Throw up G50 The movements of the hands and feet (extremities) are not active G51 Pain in the head and back G52 Baby reflex is absent G53 Cry loud and high pitch G54 Decreased appetite G55 Occurs 6-72 hours after birth G56 Respiratory disorders G57 Pale / blue baby G58 Jaundice / jaundice G59 Diarrhea G60 Bloated G61 No response G62 Release the ties or clamps of the cord that is still attached G63 Redness on the umbilical cord G64 There is a secret on the umbilical cord Yellow skin G98 High-pitched tears G99 Takipneu (breath> 60x / minute) Disease tables are used to determine diseases suffered by newborns. Disease determination is based on symptom data in new cases that are matched with symptom data in old cases that show the highest similarity. Disease data displayed in the consultation are data of disease that has the highest similarity value. , calculated based on the Sorgenfrei algorithm. In the disease table, there is code data that shows the sequence of disease codes, and newborn disease names can be seen in Table 3.

A. Results
The results of the research from the expert system with the Sorgenfrei method for diagnosing of newborn diseases can be seen in Table 4. B. Discussion The expert system uses the sorgenfrei method to diagnose newborns calculated by sorgenfrei, referred to in equation (2).
S = Value Similarity a = The same number of attributes between consultation and data stored in the database. b = The amount of data stored in the database, which is not in the consultation. c = Number of consulting data attributes, which are not stored in the database. Then the consultation is carried out by selecting the symptoms of newborn diseases, namely: 1. Chapter more than 4 times per day. 2. Unconscious.