Вы находитесь на странице: 1из 5

IMPLEMENTATION OF FUZZY C-MEANS

ALGORITHM FOR PREDICTING TRUANT

Falentino Sembiring1 , Dwi Putri sulisdianto2, Yanti Jentiner3, Erip Suratno4


1,2,3,4
Department System Information,
1,2,3,4
Nusa Putra University
1,2,3,4
Indonesia
1
Falentino.sembiring@nusaputa.ac.id, dwiputrisulisdianto@gmail.com, 3jentiner35@gmail.com, 4EripSuratno111@gmail.com
2

ABSTRACT The criteria to be assessed are There are environmental factors


Truant behavior does not merely affect student learning
that influence, individual factors
achievement, but is also classified as one form of juvenile
1.individual factors,
delinquency. Adolescents' self-esteem is a deviant behavior
2.Problems in the classroom
because there are behavioral deviations from various social
3.Teaching staff
rules or from prevailing social values and norms and become
4.School factor.
a source of problems that can endanger the upholding of the
In the research Data examined from 2018-2019 in class IX
social system. ditching activities as well as things that can set
were 323 people
the background for other problems arise. Evidently, truant
students often get involved with things that tend to harm
II. Literature Review
themselves and others such as smoking, brawls, and free
1. Maria Anistya Sasongko, Lilik Linawati, Hanna
association. The mistake of truant behavior is mostly borne by
A.Parhusip, (2015) with a thesis entitled "Application
students involved in skipping school. When case after case
of Fuzzy C-Means Algorithm to Determine the
ditching can be revealed students are the ones who become
Management of High School Students, Yogyakarta.
the burden of mistakes. This is an attitude that does not
2. Alfa Saleh, (2015) with a thesis entitled
support their potential will only add to their problems.
"Implementation of the Naïve Bayes Classification
Therefore, this research requires indicators and parameters
Method in Predicting the Amount of Use of
in finding the solution to the problem of influencing students.
Household Electricity.
For data processing, this study uses the Fuzzy C-Means
3. Novi Reandy Sasmita, Hizir Sofyan and Muhammad
method. This method was chosen because the placement of the
Subianto, (2010) with a thesis entitled "Comparison
cluster center is more appropriate than the other cluster
of Fuzzy C-Means and Fuzzy Cshell Methods using
methods.Cluster 1 is 70 students (56.4%), Cluster 2 is 25 and
Quickbird Satellite Image Data, Banda Aceh Besar
Cluster 3 is 29 students (23.4%).
4. Samuel Natalius, (2016) with a thesis entitled " Naïve
Keywords: Truant behavior, Clustering, Fuzzy c-means
Bayes Classifier Method and Its Use in Document
Classification
I . Introduction
Truant behavior habits that are often done by students will
Fuzzy C-Means FCM is a data clustering technique in which
have a negative impact on him,In addition, ditching habits can
the existence of each data point of a cluster is determined by
also reduce learning achievement. Skipping is a behavior
the value of membership. The membership value will include
caused by a lack of control of behavior. Truant behavior needs
real numbers at 0-1 intervals. FCM is one method of
to be researched to get a clear picture of the factors causing
optimizing partitioned clusters. The advantage of the FCM
the emergence of truant behavior so that appropriate action
method is that the cluster center placement is more appropriate
steps can be taken to help students to develop properly and
compared to other cluster methods. The trick is to fix the
optimally in accordance with the task of development, and
cluster center repeatedly, it will be seen that the cluster center
obtain optimal learning outcomes that can ultimately develop
will move to the right location [13]. The algorithm of fuzzy c-
their abilities and potential.
means is as follows (Yan, 1994) [14].
The decision support system created is a decision support
system that only helps in knowing what parameters dominate
Processing data Mining
students who play truant (Kegiatan Belajar Mengajar (KBM)).
Data mining processing consists of several processing
methods, namely [15]: a. The design steps in this research are:
(a) Predictive modeling which is processing data mining by b. Implement the Fuzzy C-Means algorithm.
making predictions / forecasting. The purpose of this method c. Learn how to cluster and identify the parameters
is to build a prediction model of a value that has certain used for cluster measurements.
characteristics. Examples of algorithms are Linear Regression, d. Demonstrate whether the Fuzzy C-Means
Neural Networks, Support Vector Machines, and others. algorithm can classify factors that influence student
(b) Association (Association) is a technique in data mining truancy.
that studies the relationship between data. Examples of its use
such as to analyze the behavior of students who arrive late. Data Collection and Analysis Techniques
1. Data collection technique
For example if a student has a schedule with lecturers A and
This study uses a questionnaire technique in
B, then students will arrive late. Examples of algorithms are the form of a questionnaire that is a list of
written questions addressed to data sources
FP-Growth, A Priori, and others.
and documents through written documents
(c) Clustering is a technique for grouping data into a certain needed to support the completeness of other
data
group. Examples of algorithms
2. Data analysis.
the data used is quantitative analysis.
3. Data processing
K-Means, K-Medoids, Self-Organitation Map (SOM), Fuzzy
Data processing using a Likert Scale.Likert
C-Means, and others. Example for clustering: There are five scale is a psychometric scale that is
commonly used in questionnaires and is the
islands in Indonesia: Sumatra, Kalimantan, Java, Sulawesi and
scale most widely used in research using
Papua. Then the five islands are made into three clusters based surveys
on their time: West Indonesia Time (Sumatra, Kalimantan and
Java), Central Indonesia Time (Sulawesi) and East Indonesia
Time (Papua).

(d) Classification is a technique of classifying data. The


difference. With the clustering method lies in the data, where
in the clustering the dependent variable does not exist, while
in the classification there is a dependent variable. Examples of
algorithms that use this method are ID3 and K Nearest
Neighbors
III. Research Methodology

The methodology in this study can be translated into


several steps consisting of Study of literature This helps
researchers in finding the basic theories needed in research, Figure 1. Road Mind
such as theories about data mining and Fuzzy C-Means.
Literature review This phase aims to find and collect
scientific journals, books, articles, and other literature relating
to this research as a reference. Research design

IV. The results


Grouping data according to the method used The first data is
used as data to sort problems with Fuzzy C-Means with the
SPSS program.
a. Questionnaire for factors influencing students
truant
Figure 2. factor parameters rank Student

In figure 4.1 above, it appears that after


conducting a questionnaire for each class,
the data obtained for class 9B school
factor parameters rank first, second
environmental factors for class 9A,
followed by teaching staff parameters for
class 7A. However, it needs to be
Figure 5. factor parameters rank problem
concluded that from all the existing
in class room
problems it turns out that environmental
factors ranks the most out of all the
existing problems with a value of 3.52,
second-order individual factors with a
value of 3.43 and the teaching staff
follows in the next sequence.
B. Questionnaires for factors affecting students
ditching each indicator
The following graph is displayed in the form of
the results of data analysis using a graph of each
indicator, wherein the graph explains the priority order Figure 6. factor parameters rank individual

of each indicator of each class. The following are


displayed some questionnaire charts for factors that
influence students skipping each indicator.

Figure 7. factor parameters rank school

Figure 3. factor parameters rank environment

Figure 4. factor parameters rank individual Figure 8. Comparison from other factors

Step classification predictive levels of factors that


influence students ditching with Fuzzy C-Means
1. Compilation of research data
In this study divided into two parts of research data, namely
Training Data which amounted to 199 samples and Data
Testing which amounted to 124 samples. Where Training
Data is taken from the number of students from Class 7A,
Class 7B, Class 8A, Class 8B, Class 8C, and Class 8D. While
Testing Data is taken from the number of students from Class
9A, Class 9B, Class 9C and Class 9
cluster-1 consists of 70 people, cluster-2 contains 25 people,
and cluster-3 contains 29 people who group and to learn
2. Student Data Input what variables are included in the category of each cluster

This student data is inputted to SPSS for clustering.


The learning process includes data with SPSS.

Figure 10 Final Cluster Centers

Figure 9 Final Cluster Centers

System implementation

For the implementation of the author using Microsoft Visual


Basic 2010. Guidance Counseling (BK) teacher input data as
parameters that have been obtained from the questionnaire
and will be analyzed using the Fuzzy C-Means method so that
data can be displayed on the system. As for what was made
can be displayed in Figure 9.

In table 4.2 above is the result of a


questionnaire conducted on 5 observers who played a
part of the user and was taken at random.
Figure 9 The Results Evaluasi SQA Skor = <81.6>*0.125 + <86.6>*0.125 + <86.4>*0.125 +
<81>*0.125 + <86>*0.125 + <91.4>*0.125 + <88.6>*0.125 +
<87.2>*0.125

The resulting average score is 86.1, while


the optimal value for a software that meets quality
standards based on the SQA test is 86.1
successfully implemented and has been
V. Conclusion proven at the research testing stage.

3. the system created produced a 86.1%


1. Research that has been done on the data of
questionnaire from the Metric of
students of Junior High School with the
Software Quality Assurance (SQA) test.
factors that influence students who play
truant produced the following things: 4. Cluster 1 with 70 students divided by
124 students multiplied by 100%
2. The use of the Fuzzy C-Means method in
(56.4%), Cluster 2 with 25 students
this study is able to provide the decision
divided by 124 multiplied by 100%
of students who fall into groups of
(20.2%) and Cluster 3 with 29 students
cluster 1, cluster 2, cluster 3 and cluster
divided by 124 multiplied by 100%
4. This proves that the Fuzzy C-Means
(23.4) %) results from SPSS
method applied in the system has been

REFERENCES [8] KingOFMath, “Metode Clustering dan Fuzzy C-

1] Harianto, Mamak, “Pengaruh Konseling Kelompok Means”, Jakarta, 2015.

Terhadap Penanganan Siswa Membolos pada Kelas [9] Situngkir, Nurmaita, “Sistem Pendukung Keputusan

VIII di MTs. Nurul Huda Sedati Sidoarjo”, Sidoarjo. Seleksi Pemberian Beras Miskin (Raskin) Dengan
2016 Metode K-Nearset Neighbour (K-NN) Pada

[2] Patil, T. R., Sherekar, M. S., 2013, Performance Kecamatan Medan Tuntungan Berbasis Client

Analysis of Naive Bayes and J48 Classification Server”, Medan: 2015.Han, J. dan M. Kamber. 2006.

Algorithm for Data Classification, International Data Mining: Concepts and Techniques, Second
Journal of Computer Science and Applications, Vol. Edition. Morgan Kaufmann Publishers. San
6, No. 2, Hal 256-261 Francisco.

[3] Bustami., 2013, Penerapan Algoritma Naive Bayes [10] Maria Anistya S et al, “Penerapan Algoritma Fuzzy
Untuk Mengklasifikasi Data Nasabah Asuransi, C-Means Guna Penentuan Penjurusan Program

TECHSI : Jurnal Penelitian Teknik Informatika, Vol. Peserta Didik Tingkat SMA”, Yogyakarta. 2015

3, No.2, Hal. 127-146. [11] Novi Reandy et al,“Perbandingan Metode Fuzzy C-

[4] Ridwan, M., Suyono, H., Sarosa, M., 2013, Means dan Fuzzy Cshell menggunakan data Citra

Penerapan Data Mining untuk Evaluasi Kinerja Satelit Quickbird”, Banda Aceh Besar. 2010

Akademik Mahasiswa Menggunakan Algoritma [12] http://kbbi.web.id//terap-2


Naive Bayes Classifier, Jurnal EECCIS, Vol 1, No. 7, [13] Nurhikmah Megawati, Moch. Abdul Mukid, and Rita
Hal. 59-64 Rahmawati, "Segmentasi Pasar Pada Pusat
Perbelanjaan Menggunakan Fuzzy CMeans (Studi
[5] Cahyo Darujati, Agustinus Bimo Gumelar,
Kasus : Rita Pasaraya Cilacap)," Jurnal
“Pemanfataan Teknik Supervised untuk Klasifikasi
Gaussian, vol. 2, no. 4, pp. 343-350, 2013
Teks Bahasa Indonesia, Surabya: 2012
[6]http://www,ftsm.ukm.my/irpa/EA012/bunggul%20paper2/F [14] Yohana Nugraheni, "Data Mining dengan Metode
ull012.pdf Fuzzy untuk Customer q Relationship Management
Geophysical research Abstracts, Vol. 7, 11076, 2005 (CRM) pada Perusahaan Ritel," Universitas Udayana,
SRef-ID: 1607-7962/gra/EGU05-A-11076 European Denpasar: Thesis 2011
Geosciences Union 2005 (diakses pada tanggal 20 [15] Larose, 2005, “Discovering Knowledge in Data: An
Oktober 2018) Introduction to Data Mining” John Willey &
[7] Nur Fatihah Aziizatul Munawaroh,“Fuzzy, Sons, Inc.
Kekurangan, Kelebihan, Logika”, Semarang: 2015.
[16] Software Quality Assurance (SQA)

Вам также может понравиться