227 EXPERT SYSTEM DEVELOPMENT TO IDENTIFY EMPLOYEE PERSONALITY TYPES USING DEMPSTER SHAFER THEORY Julia Fajaryanti1*), Rogayah2 Informatika, Fakultas Teknologi Industri Universitas Gunadarma www.gunadarma.ac.id julia@staff.gunadarma.ac.id, rogayah@staff.gunadarma.ac.id (*) Corresponding Author Abstrak Sumber daya manuisa menjadi aset penting bagi perusahaan untuk berkembang dan mewujudkan cita -cita perusahaan. Salah satu usaha untuk mengoptimalkan kapasitas karyawan adalah dengan mengetahui kepribadiannya. Kepribadian merupakan bentuk yang dimiliki seseorang individu dalam bertingkahlaku dan segala watak yang membedakan antara seorang individu yang satu dengan seorang individu lainnya. Mengetahui kepribadian karyawan menjadi suatu hal yang penting bagi perusahaan dan karyawan itu sendiri. Karena dengan mengetahui kepribadian seseorang perusahaan dapat dapat memaksimalkan potensi karyawan dan dapat menepatkan posisi tertentu yang sesuai dengan kepribadian karyawan. Penelitian ini bertujuan untuk mengimplementasikani dempster-shafer theory pada mesin inferensi dalam membangun sistem pakar untuk mengidentifikasi tipe kepribadian karyawan. Dempster-shafer theory dapat melakukan perhitungan probabilitas sehingga dapat dilakukan pembuktian berdasarkan tingkat kepercayaan dan penalaran yang logis. Sistem yang dikembangkan mampu mengidentifikasi tipe kepribadian karyawan melalui sifat atau gejala yang ada pada diri karyawan. Selain itu, sistem dapat menampilkan hasil diagnosis dengan penjelasan tentang tipe kepribadian, sifatnya dalam pekerjaan dan pekerjaan atau posisi yang cocok untuk tipe kepribadian tersebut. Berdasarkan hasil uji akurasi yang diperoleh dari hasil perbandingan diagnosa sistem pakar dengan analisis seorang pakar menunjukkan nilai akurasi mencapai 85%. Kata kunci: dempster-shafer theory; sistem pakar; mesin inferensi; tipe kepribadian Abstract Human resources are an essential asset for the company to develop and realize the company's goals. One of the efforts to optimize the capacity of employees is to know their personality. Personality is the form an individual possesses in behaving and all the characteristics that distinguish one individual from another. Understanding employees' personality is essential for the company and the employees themselves. Because by knowing a person's personality, the company can maximize the potential of employees and place certain positions that suit the employee's personality. This study aims to implement the dempster-Shafer theory on an inference engine in building an expert system to identify employee personality types. Dempster- Shafer's approach can perform probability calculations so that evidence can be carried out based on confidence and logical reasoning. The system developed can identify the employee's personality type through the nature or symptoms that exist in the employee. In addition, the system can display the diagnosis results with an explanation of the personality type, its character in work and occupations or positions suitable for that personality type. Based on the results of the accuracy-test obtained from the comparison of expert system diagnoses with the analysis of an expert, the accuracy value reaches 85%. Keywords: dempster-Shafer theory; expert system; inference engine; personality type INTRODUCTION Human resources are one of the most important elements for a company or organization to run well. Where human resources have a significant influence on the success of achieving goals in order to realize the company's vision and mission. Today, many companies have viewed human resources not only as resources but rather as valuable capital or assets that must be cared for and maintained for development (Zulkarnaen, 2018). One of the developments of human resources is knowing each employee's personality. The reason is that each employee has various 228 psychological behaviours that must be processed to achieve company goals. Personality is the form an individual possesses in behaving and all the characteristics that distinguish one individual from another (Sya'baniah et al., 2019). Personality is something that someone needs to know so that everyone can develop the potential that exists in each individual (Darmansah et al., 2021). Human personality types have been studied and summarized into 4 (four) types. The four types are included in the proto-psychological theory. This theory was first discovered in the century BC by Hippocrates, then by Galen was developed into a medical approach. According to this theory, human personality is grouped into four categories: choleric, sanguine, phlegmatic, and melancholic. Knowing employees' personalities is essential for the company and the employees themselves. Because by knowing a person's personality, the company can place certain positions that are appropriate and can maximize the potential of employees. To find out a person's personality type, you can use technology such as an expert system. Expert system is a sub-field of artificial intelligence that can manage and draw conclusions based on specific rules obtained from knowledge (Borman et al., 2020). Expert systems are also called knowledge-based systems, and this is because the expert system provides a collection of knowledge obtained from an expert and the required knowledge sources (Putri, 2018). The purpose of developing an expert system is to build a system that can ease human work, especially those related to the use of the ability and experience of an expert (Sucipto et al., 2019). One of the inference engine methods in expert systems that can overcome uncertainty is the dempster-Shafer theory. The dempster-Shafer idea comes up with an approach to calculating probabilities so that proof can be done based on the level of belief and logical reasoning so that it can be used in combining information (evidence) (Rahmanita et al., 2019). Several studies have shown that the dempster-Shafer theory can reasonably produce expert systems. Research the expert system used to diagnose gastric disease using the dumpster-Shafer algorithm (Ardiansyah et al., 2019). In this study, the dempster Shafer produced a system based on the confidence function with an accuracy of 95%. Further research, developing a system that can diagnose human skin diseases using the Dempster- Shafer algorithm (MZ et al., 2020). In this study, the expert system developed for each symptom has a confidence value used to calculate the density that results in conclusions. Based on the accuracy test, it produces a discount of 90%. Meanwhile, the Mean Opinion Score (MOS) test resulted in a MOS size of 4.35, which means the system has a good feasibility level. Furthermore, research on developing an expert system for diagnosing oral cancer shows that the Dempster- Shafer algorithm can overcome the uncertainty in constructing an inference engine. It is indicated by the results of the accuracy test of 86.6% (Napianto et al., 2018). This study aims to implement the dempster-Shafer theory on the inference engine in building an expert system to identify the employee personality type. An expert system is built based on a website to make it easier for users to use the system. The system can recognize personality types based on symptoms or a person's character. The system also includes an explanation of the personality type, its nature in the job, and the job or position that is suitable for that personality type. RESEARCH METHODS Research needs to be arranged in stages so that the research carried out is by the objectives to be achieved. The locations of research carried out in this study are presented in Figure 1. Figure 1. Research Stages 229 Identification of problems The first step is identifying the main problem to get the right solution. The output of this stage is a statement of the problem to be solved. The situation in the world of work that is often experienced is the inaccuracy of a person's position or job with his personality. It results in not working optimally. So we need a system that can identify a person's personality type to suit his career. Knowledge Acquisition Expert systems are also known as knowledge-based systems, so they cannot be separated from the collection of knowledge from experts or experts. To get an understanding from an expert, a knowledge acquisition stage is needed. Knowledge acquisition is the process of extracting, structuring, and organizing knowledge from knowledge sources, so that knowledge can become a knowledge base that is the basis for decisions in expert systems (Anita et al., 2019). The output of this stage is knowledge in the form of symptom data, personality type data, and diagnostic rules. This study was obtained through interviews and gathering sources of knowledge through books to get the required data, such as data on symptoms, personality types, and the level of confidence for each sign and personality type. The data was obtained from the results of consultations with experts or experts, namely a psychologist. From the results of interviews with psychologists and collecting data from books, he got 4 (four) personality types with 30 characters or symptoms. The personality type used is based on Hippocratic Theory and developed by Galen. According to this theory, human personality is grouped into four categories: Choleric, Sanguine, Phlegmatic, and Melancholic. Knowledge Representation The next stage is knowledge representation, where the results obtained in knowledge acquisition will be organized regularly to encode expert knowledge into appropriate media forms. Knowledge representation is vital in developing expert systems because a good solution will also depend on a good word. If the knowledge representation is not made correctly, the impact will affect the next stage, and the resulting system is not as desired (Nasution & Khairuna, 2017). From the knowledge acquisition process, knowledge is obtained that will be used as a knowledge base, then knowledge representation is carried out using a decision table, and rules are made based on the understanding that has been obtained from experts, which will later be used to build an inference engine. Inference Engine The inference engine can be said to be part of an expert system that carries out a reasoning function that utilizes rules with specific patterns (Annisa, 2018). The inference engine will perform a search based on the rules in the knowledge base, which results from the conclusion in the form of a solution that suits your needs. In this study, the inference engine used is the dempster-Shafer theory. Dempster-Shafer's approach performs probability calculations to obtain evidence based on the level of belief and logical reasoning because later, it will be used to combine evidence to get firm conclusions. Dempster-Shafer's theory generally uses the [Belief, Plausibility] format. Bel (Belief) is a measure of evidence's ability to support a condition (Rahmanita et al., 2019). If belief has a value of 0, then the indication is that there is no evidence. On the contrary, there is a certainty if it has a value of 1. The belief function can be denoted by equation (1). )()( YmXBel X .................................................................. (1) Where Bel (X) is a belief of (X), while m (Y) is a mass function of (Y), plausibility (Pls) is denoted in equation (2) below. )'()'()( XmXBelXPls n XY  11 .................................. (2) Bel(X) is a belief of (X), while m is a mass function. Pls(X) is the plausibility of (X). Plausibility can have a value between 0 to 1, if there is belief in X' then Belief (X') = 1 which results in the result Pls(X) = 0. Table 1 below is the possibilities that occur between belief and plausibility. Table 1. Possible Bell and Pls Ranges Certainty Description [1, 1] [0, 0] [0, 1] [Bel, 1] where 0< Bel<1 [0, Pls] where 0