Implementation Of the Mamdani Fuzzy Method to Evaluate the Performance of Lecturers in The Research Field

— Today, information technology, especially soft computing technology, has grown very rapidly. One of the soft computing technologies that has been widely developed is fuzzy logic, because it can be used to measure various phenomena that are ambiguous, disguised or fuzzy. One of the research topics that uses the application of fuzzy logic is the assessment system in the field of research. Research by Lecturers at the University of Graha Nusantara Padangsidimpuan in Simlitabmas Data Still in the Guidance Category to upgrade to the Middle Category UGN Padangsidimpuan lecturers are challenged to be able to develop, devote, and apply the knowledge needed in research. For that we need an application that can be used to calculate and record the performance of Lecturers on the resulting Research. The purpose of this study is to apply fuzzy logic with the Mamdani method in assessing the research performance of lecturers at the University of Graha Nusantara Padangsidimpuan. This research uses Mamdani Fuzzy Logic. Fuzzy Mamdani method is a way to map an input space into an output space. This method is a mathematical framework used to represent uncertainty, ambiguity, imprecision, lack of information, and partial truth. The stages of research using the Mamdani method are Creating Input Variables taken from Sinta-accredited Articles, Simlitabmas Grant Articles and Articles in International and National Journals. Finding the Max-Min Value of Each Variable. Creating a fuzzy set using the Mamdani method. Creating Assertions with Defuzzification using Matlab.


I. INTRODUCTION
According to Junaidi et al (2020), teaching staff is an important component to enter high-quality lecturers, therefore to become teaching staff must have high abilities in accordance with these limitations (Wawan et al., 2021). Meanwhile, Universitas Graha Nusantara (UGN) Padangsidimpuan has a vision of becoming a superior, independent and nationally competitive higher lecturer by 2030 (Rustum et al., 2020). To realize this vision, UGN has one of its missions, namely Organizing education, research and community service, as well as conducting studies and periodic studies. (Sadi, 2020)  a. An ascending linear representation that represents the increase in the set, starting from the domain value which has zero membership degree (Wu & Xu, 2021), moving to the right to the domain value which has the same higher degree of membership. The steps carried out in this study are shown in Figure 3 below.
Based on the research steps in Figure 3, each step can be explained as follows: 1. Data collection The data needed for this research is research data from the Faculty of Graha Nusantara University in 2020 and 2021 which is taken from data from the Sinta cluster, Simlitabmas and journal publications. 2. Data identification Data identification is done to select the variables needed to perform calculations and analyze the problem.
3. Data processing The data processing stage is to create a Mamdani fuzzy system using Matlab software. 4. System test At the system testing stage, testing and simulation will be carried out to evaluate lecturer research activities using fuzzy logic.
III. RESULTS AND DISCUSSION This study examines the Mamdani fuzzy method in evaluating research activities of the Faculty of Graha Nusantara Padangsidimpuan University using an application built with Matlab R2013a software. In this study, it consists of 3 input variables, namely the variables that are used as evaluation materials, which include variables from Sinta, Simlitabmas and Klater Jurnal, which will provide 33 rules or 27 rules. The fuzzy set for the input and output variables is presented in Table 1 below:  In this study, the discussion process was carried out in several stages, namely: 1. Determine the input variables taken from the lecturer's research assessment data, where the variables used are the sinta variable, the simlitabmas variable, and the journal cluster variable. 2. Fuzzification: determining the degree of membership of the input and output variables.

Fuzzy logic operations must be performed
if the previous part of more than one statement performs fuzzy logic operations. The final result of this operation is the degree of truth of the antecedent, which is a single number. Fuzzy operators to perform operations and and or can be made independently. 4. Implication: Apply the implication method to determine the final form of fuzzy set output. The consequence or inference of a fuzzy rule is determined by filling the output of the fuzzy set with the output variable. The implication function used is Min. 5. Aggregation: The process of combining the outputs of all if-then rules into one fuzzy set using the Max function. 6. DefuzzificationThe inference process in the application of fuzzy statements uses the MIN implication function. In addition, the composition of all fuzzy outputs is done using max. Then do validation or called defuzzification using Centroid. In this method, a crisp solution is obtained by taking the center point of the fuzzy area as follows :

Fuzzy Statement Analysis Using Matlab
The validation of research evaluation data for graha nusanatara university lecturers using the mamadani method can also be done using the matlab fuzzy toolkit version R2013a. This software serves to interpret the variables of lecturer research activities This study has 3 input variables and 2 output variables. The input variables consist of Sinta, Simlitabmas, and cluster. 2020 and 2021. While the minor and interim release variables. This can be seen in Figure 4. Defuzzification fills the output variable with one number using the centroid or area center method. The last step in this implementation is the process of taking the input value to get the output value. In this study, the input value is 104, the initial output value is 103, and the final stage output is 324.

Measurement accuracy rate
The definition of accuracy is how close the measurement result is to the actual number. Because this study is so precise, we start with the number of measurements, the Y value of the Mamdani method, which uses a standard set of values to give the correct result. The default value of the Mamdani method is the value of the output variable for assessing lecturer research activities, determined using the membership function.