Corresponding Author: Ade Bani Riyan
Email: [email protected]
INTRODUCTION
The student point system is an application program for
recording student achievement points and violations (Dwijaya, 2020). Not recording and disseminating information on
achievement results makes students less motivated to improve. Data on student
achievement and violations is needed, especially for scholarship distribution (Rachmawaty, 2016). Having student achievement data will make it easier
to distribute accurate scholarships. To improve student achievement, an
application program is necessary as a tool for recording and disseminating data
on student achievement information in real time, accurately and effectively.
Researchers used the Clustering Method in calculating data to determine the
accuracy of scholarship distribution for outstanding students.
Education is one indicator of whether a country is
progressing or not (Al-Hendawi et al.,
2023). If the government does not pay attention to the
progress of its education, entry will not move. Education is the main thing in
a country because it greatly influences other fields (Muhardi, 2004). Such as government, finance, defense and so on. In
the continuity of a country in the future, education is also very important
because the education of the next children of a country that is not considered
will threaten the sustainability of the country in the future (Ramadhanu et al.,
2021).
Data mining �is a process of artificial intelligence, machine learning,
statistics, and mathematics to extract and identify useful and related
information from large databases (Gorunescu, 2011).
The clustering method groups several data or objects
into groups (groups) so that each group contains data that is as similar as
possible and different from data/objects in other groups (Maukar et al., 2022). While cluster analysis, according to Eka Haryati in
his research was used to determine patterns with high characteristics (Haryati et al., 2022).
The clustering method has so far been applied in
various fields, as written in research and journals (Windarto et al., 2017).
�Clustering is the process of grouping many data points into
two or more groups so that data points belonging to the same group are more like
one another than in different groups, based solely on the information available
with the data points (Nidheesh et al., 2017).
KDD is a method used to search for knowledge from a
database. In his research �explains that the results of knowledge can be used as a
knowledge base that can be used to make a decision (Adiya & Desnelita,
2019). In more detail, the KKD
process is in the following figure adopted from (Gullo,
n.d.).
Figure
1. Steps of the KKD Process
The K-Means algorithm is one of the algorithms often
applied in grouping because of its efficiency and simplicity (Harding et al., 2006). It is recognized as part of the top 10 data mining
algorithms by IEEE (Wu et al., 2008).
Based on the above background, the purpose of this
study is to find out and analyze the design of the student point system to
improve student achievement using the clustering method. With this research, it
is hoped that it can be a solution for recording and disseminating student
achievement information, so that students are more motivated to excel and
increase the level of accuracy in distributing scholarships for outstanding students.
METHODS
In this study, researchers used data
from point system applications that run-in schools. This study has four stages
illustrated in Figure 2. Research Stages.
Figure 2. Research Stages
1)
Study
Literature
The literature
study was carried out by collecting various methods and theories related to the
problems in the research conducted, namely the use of the K-Means clustering
algorithm. Literature studies are obtained from multiple sources, including
magazines, articles or scientific papers that are used to strengthen the
theoretical basis in research. Several journal references were used as an
overview in this study, namely research conducted by " Interactive Web-Based Smart School at Amaliyah Private
Elementary School Sunggal with the K-Means Cluster Algorithm Information
Systems Study Program" and "Implementation of the K-Means Method in
Mapping Student Groups Through Lecture Activity Data" (Fauzi &
Samsudin, 2022); (Rosmini
et al. , 2018). �
2)
Data
collection
The stages of data
collection are the process of collecting data through the student point system
application database from July 2021 to December 2021. The total data obtained
from this period amounted to 537 rows.
3)
Processing
Data
The data obtained
at the data collection stage is carried out by the next process, namely the
pre-processing step. At this stage several activities are carried out, namely:
a)
Data
selection was carried out to collect data that is suitable for analysis
purposes, namely selecting data with the characteristics of Level, Class,
Achievement Points, Violation Points;
b)
Convert
data into a form that is more suitable for analysis, namely converting classes
into integer form (numbers) to facilitate research and as a requirement for
data to be read by the rapid miner tool;
c)
Clearing
data is removing some inconsistent data from the collected data. Some of the
data characteristics of the data obtained after pre-processing the data can be
seen in Table 1 Sample dataset, as follows:
Table
1. Sample Dataset
|
Name
|
Gender
|
Level
|
Class
|
Achievement Points
|
Violation Points
|
|
Sister Al Aina
|
P
|
3
|
10
|
50
|
25
|
|
Ajat Sudrajat
|
L
|
3
|
10
|
25
|
30
|
|
Aji Soleman
|
L
|
3
|
11
|
30
|
75
|
|
Alika Trista Aulia
|
P
|
3
|
10
|
25
|
30
|
|
Amanda
|
P
|
3
|
10
|
75
|
20
|
|
Amelia Putri
|
P
|
3
|
11
|
20
|
20
|
|
Andr� Afrijal Maolana
|
L
|
3
|
10
|
30
|
25
|
Description Level
1 = Elementary School
2 = Junior High School
3 = Senior High School
RESULTS AND DISCUSSION
537 data were processed using two methods. First the data
is loaded into Rapid Miner and run using the Elbow method to get the right number of clusters before starting the
clustering process. Then, the data is processed using the K
-means way to get clustering results. An overview of the
elbow method process can be seen in
Figure 3. The Elbow Method process and an overview of the
k-means method process can be seen in Figure 4. The K-Means Method Process.
Figure
3. Elbow Method Process
Figure
4. The process of the K-means method
Experimental results k-2
to k-10 and seed value = 10. Seed is a random number in cluster generation with
seed value 10 as the default number used as a process reference. This
normalization produces an output value between 0 and 1. Then the process of
grouping the datasets into their respective groups based on the similarity of
characteristics is done by calculating the distance value using the Euclidean
Distance in the equation and the K-Means algorithm for processing orders. Then the
cluster results were analyzed and evaluated to find the optimal number of K
using the Elbow method. The Elbow method calculates the
largest SSE depreciation difference and is in the shape of an elbow.
Calculation of SSE using the equation. After the clustering trial process on
the dataset, the data processing is carried out by calculating the distance
value to determine the number of clusters, the results are shown in Table 2�comparison
of Average Results.
Table 2. Comparison of Average
Results
|
Clusters
|
Average
|
|
K2
|
385,788
|
|
K3
|
69,047
|
|
K4
|
36,829
|
|
K5
|
28,161
|
|
K6
|
20,936
|
|
K7
|
16.120
|
|
K8
|
12.172
|
|
K9
|
10,789
|
|
K10
|
9,348
|
Figure
5. Graph of Clustering Results
We try to do a test with the Elbow method to find out how
many clusters are suitable for the analysis process. Based on the test results in
Table 2. Comparison of Average Results, we visualize the table in graphical
form and look for clusters with the most angular lines. Based on Figure 5 it
can be seen that point K4 shows the most angular results compared to other
points. So it can be concluded that the most optimal cluster is according to
the elbow method using 4 groups. Furthermore, testing was carried out again on
the rapid miner using the recommended clustering, the results of the test using
the K-Means algorithm with 4 sets can be seen in Figure 6. Results of 4 Cluster
Models.
Figure
6. Results of 4 Cluster Models
The results of clustering with 4 clusters show that the lowest achievement point
values are in cluster 4 and the highest in cluster 3. Detailed clustering comparisons can be seen in Table 3�results of the Clustering Process.
Table
3. Clustering Process Results
|
attributes
|
Cluster_0
|
Cluster_1
|
Cluster_2
|
Cluster_3
|
|
Level
|
1,725
|
1690
|
1,797
|
1676
|
|
Class
|
5514
|
5,561
|
6004
|
5,595
|
|
Achievement Points
|
75
|
25,032
|
25,254
|
50
|
|
Violation Points
|
26.147
|
75
|
24,831
|
25
|
CONCLUSION
The results of the clustering research
on student data points obtained clustering with the most achievement points,
namely clustering 2, with a total of 25,254 Achievement Points. The total
number at cluster level 2 is 1,797, where these results show the number is
close to 2,000 or 2, which is the result of data transformation from the junior
high school level raised in Table 1. Sample Dataset. The foundation, as an
institution that oversees schools, can see achievement data from the school
level, making it easier to allocate scholarships for outstanding students based
on their school level.
REFERENCES
Adiya, M. H., & Desnelita, Y.
(2019). Jurnal Nasional Teknologi dan Sistem Informasi Penerapan Algoritma
K-Means Untuk Clustering Data Obat-Obatan Pada RSUD Pekanbaru. Vol, 1,
17�24.
Al-Hendawi, M., Keller, C., & Khair,
M. S. (2023). Special Education in
the Arab Gulf Countries: An Analysis of Ideals and Realities. International
Journal of Educational Research Open, 4, 100217. https://doi.org/10.1016/j.ijedro.2022.100217
Dwijaya, D. A. (2020). Perancangan
Aplikasi Untuk Pelanggaran Dan Prestasi Siswa Pada Smp Kartika Ii-2 Bandar
Lampung. Jurnal Informatika
Dan Rekayasa Perangkat Lunak, 1
(2), 127�136. https://doi.org/10.33365/jatika.v1i2.313
Fauzi, M. S., & Samsudin, S. (2022). Smart School
Berbasis Web Interaktif di SD Swasta Amaliyah Sunggal dengan Algoritma K-Means
Cluster. Jurnal Sisfokom (Sistem Informasi Dan Komputer), 11(3),
332�341.
Gorunescu, F. (2011). Data Mining: Concepts, models and
techniques (Vol. 12). Springer Science & Business Media.
Gullo, F. (n.d.). From patterns in data to knowledge
discovery: what data mining can do. Phys. Procedia 62, 18�22 (2015). 3rd
International Conference Frontiers in Diagnostic Technologies.
Harding, J. A., Shahbaz, M., & Kusiak, A. (2006). Data
mining in manufacturing: a review.
Haryati, A. E., Wijaya, T. T., Wen, G. K., & Thobirin, A.
(2022). Fuzzy subtractive clustering (FSC) with exponential membership function
for heart failure disease clustering. International Journal of Artificial
Intelligence Research, �6(1). https://doi.org/10.29099/ijair.v7i1.306
Maukar, A. L., Marisa, F., & Widodo, A. A. (2022). Analisis
Data Penerimaan Mahasiswa Baru Berbasis K-Means. JIKO (Jurnal Informatika
Dan Komputer), 6(2), 142�147.
Muhardi, M. (2004). Kontribusi pendidikan dalam meningkatkan
kualitas bangsa Indonesia. Mimbar: Jurnal Sosial Dan Pembangunan, 20(4), 478�492. ttps://doi.org/10.29313/mimbar.v20i4.153
Nidheesh, N., Nazeer, K. A. A., & Ameer, P. M. (2017). An
enhanced deterministic K-Means clustering algorithm for cancer subtype
prediction from gene expression data. Computers in Biology and Medicine,
91, 213�221. https://doi.org/10.1016/j.compbiomed.2017.10.014
Rachmawaty, D. T. (2016). Pengaruh beasiswa Bidikmisi
terhadap prestasi belajar mahasiswa penerima beasiswa Bidikmisi di UIN Syarif
Hidayatullah Jakarta.
Ramadhanu, A., Defit, S., & Kareem, S. W. (2021). Hybrid
Data Mining with the Combination of K-Means Algorithm and C4. 5 to Predict
Student Achievement. International Journal of Artificial Intelligence
Research, 5 (2), 180�189.
Rosmini, R., Fadlil, A., & Sunardi, S. (2018).
Implementasi Metode K-Means Dalam Pemetaan Kelompok Mahasiswa Melalui Data
Aktivitas Kuliah. IT Journal Research and Development, 3 (1),
22�31.
Windarto, A. P., Komputer, I., & Bangsa, T. (2017).
Implementation of Data Mining on Rice Imports by Major Country of Origin
Implementation of Data Mining on Rice Imports by Major Country of Origin Using
Algorithm Using K-Means Clustering Method. No. November. https://doi.org/10.29099/ijair.v1i2.17
Wu, X., Kumar, V., Ross Quinlan, J., Ghosh, J., Yang, Q.,
Motoda, H., McLachlan, G. J., Ng, A., Liu, B., & Yu, P. S. (2008). Top 10
algorithms in data mining. Knowledge and Information Systems, 14 (1),
1�37. https://doi.org/10.1007/s10115-007-0114-2
