STUDENT ACHIEVEMENT CATEGORY WITH K-MEANS AND K-NEAREST NEIGHBORS

,


INTRODUCTION
Based on the decision of the Director General of Higher Education, Ministry of Education and Culture of the Republic of Indonesia Number 84/E/KPT/2020 concerning Guidelines for Implementing Compulsory Courses in the Higher Education Curriculum, it was decided that every college curriculum must contain courses on religion, Pancasila, citizenship and Indonesian.This course is mandatory because it has the function of forming the character and civility of students with dignity.The minimum study load for each course is 2 semester credit units.These four courses were developed by every university in Indonesia.The importance of these four courses is that the government under the Director of Learning and Student Affairs provides assistance in developing mandatory curriculum courses (MKWK) to develop MKWK tools and conduct MKWK implementation surveys at each university.Apart from MKWK at Cirebon Muhammadiyah University there are university courses.
One of the subjects that is becoming common at universities is mathematics.
Mathematics is a mandatory university subject because mathematics is seen to play a large role in everyday life.Typical mathematical thinking patterns such as systematic, scientific, logical and critical make mathematics courses a provision for students in decision making.
Low achievement in learning mathematics not only has an impact on academic achievement but also affects cognitive development and logical thinking.Therefore, various efforts are made by lecturers to be able to improve student achievement.One of them is by knowing the student Achievement Index which is used to measure the ability to think mathematical reasoning because mathematical reasoning can contribute to student learning achievement as done by (Gde Somatanaya, 2017).
This research uses mathematics course grades because apart from the objectives of the MKWK, mathematics influences other achievements.This can be seen from the bibliometric results obtained based on a search via Scopus with the keyword achievement during 2018 to 2023, 2782 articles were obtained with the theme of learning achievement.Visualization can be shown in the following image.K-Mean and K-NN are widely used in various studies in education.One of them is for the classification of bidikmisi recipient candidates by using the K-NN algorithm, sentiment (Hilda Nur Zerlinda et al., 2019), analysis of the 2012 curriculum on twitter social media using the K-Nearest Neighbor method and feature Selection Query Expansion Ranking (Mentari et al., 2018), and Application of K-Nearest Neighbor for Classification of Graduation Rates in Students (Purwaningsih & Nurelasari, 2021).This shows that the KNN and K Mean methods can be used in education.

METHODS
The research was conducted at the Muhammadiyah University of Cirebon.Research data was obtained from the academic system owned by the Cirebon Muhammadiyah University called the great system.Data collection was carried out on all study programs for level 1 students for Pancasila, mathematics, Indonesian and religion courses.However, at the tabulation stage, complete data was only available for 7 study programs for each student.So the data processed is the final student grades from 7 study programs for the same student.
The 7 study programs are Nutrition, Government Science, Mathematics, Early Childhood Education, PGSD, Industrial Engineering and Information Engineering.
The data obtained used secondary data originating from the academic system at Muhammadiyah University of Cirebon.This system is called a great system.The data taken is data from MKWK and Mathematics course grades.The MKWK data that can be downloaded are those from which the data is taken, namely Mathematics, Indonesian, Religion and Pancasila.Citizenship courses cannot be downloaded because there is a system migration from SIMAKU to the great system.
The number of students studied was 369 students.The scores for four courses were taken from 369 students.Then data processing was carried out using K-Means and KNN.The achievement categories that will be achieved are high, medium and low categories.
The data processing technique in this research uses K-Mean and K-NN calculations using orange software.The stages used are uploading files, uploading data, carrying out prepossessing such as avoiding outliner data and carrying out normality tests after that processing the data using K-mean.After completing the K-mean steps, the K-NN process is carried out.The stages carried out are the same as in K-Means processing, the data uploaded is based on the K-Means results obtained.
The data processing technique in this study uses K-Mean and K-NN calculations using orange software.(Purnamasari & Widiastuti, 2017) states that K-Means is one of the unsupervised learning methods.The K-Means method is included in the non-hierarchical clustering category which is easy and simple, finds convergence quickly, and can adapt to data distribution.The research data is unsupervised data that needs to be clustered so it requires the K-Means method.K-Means is used to determine the level of learning achievement based on the scores obtained by students in mathematics courses.(Cholil et al., 2021) stated that the KNN algorithm has several advantages, including robustness against training data that has a lot of noise and a large amount of data.In this study, the data taken is secondary data consisting of hundreds of data so that it requires data processing with KNN.The combination of both can produce good data interpretation as done by (Aditra Pradnyana et al., n.d )on the division of college classes based on learning quality.The data obtained can clearly map the results of the learning process carried out.In this study, the stages used are uploading files, Prima: Jurnal Pendidikan Matematika ◼ 253

Student Achievement Category With K-Means And K-Nearest Neighbors
Lusiyana, Dasari uploading data, doing prepossessing such as avoiding outliner data and doing normality tests after that processing data using K-mean.After the stages in K-mean are completed, the K-NN process is carried out.The stages carried out are the same as in K-Means processing, the uploaded data is based on the results obtained by K-Means.

RESULTS AND DISCUSSION
Results of data processing using orange software on secondary data with a total of 369 values at the Muhammadiyah University of Cirebon.The course scores used are mathematics, Pancasila, religion and Indonesian in 7 study programs, namely the 7 study programs, namely Nutrition, Government Science, Mathematics, PAUD, PGSD, Industrial Engineering and Information Engineering.The data tabulation is as follows.the data tabulation results above were then processed using orange software.the steps for using preprocessing are as follows.
Achievement Category With K-Means And K-Nearest NeighborsLusiyana, Dasari

Figure 1 .
Figure 1.Bibliometric student achievment There are five clusters based on bibliometric analysis using Vos Viewer.The largest cluster is the red cluster.The graph shows that academic achievement is closely related to mathematics.Because the courses used other than MKWK are mathematics.The focus of this research is what categories of academic achievement are based on MKWK and Mathematics subjects using K-Mean and K-NN.

Figure 2 .
Figure 2. Preprosessing dataIn Figure1is the stage to process data using orange software by inputting data in the file.Then do the colom selection and preprocessing process.The results of data preprocessing obtained the following data

Figure 5 .
Figure 5. Sample result K-Means the second data processing was carried out using k-nn.based on the results of k-means, three categories were obtained: c1 in the high category, c2 in the medium category and c3 in the low category.this data can be presented in the following graphical form.

Figure 6 .
Figure 6.Distribution categories Based on the figure above, 3 categorizations are obtained, namely C1, C2 and C3.C1shows the highest number and the lowest number is obtained by C3.