JURNAL RISET INFORMATIKA
Vol. 2, No. 3 June 2020
P-ISSN: 2656-1743 |E-ISSN: 2656-1735
DOI: https://doi.org/10.34288/jri.v2i3.137
101
The work is distributed under the Creative Commons Attribution-NonCommercial 4.0 International License
ANALYSIS OF CUSTOMER SATISFACTION ON SERVICE ARTHASPA
SERVICE WITH APPROACH ALGORITHM C4.5
Sofian Wira Hadi1*, Wawan Kurniawan2, Kudiantoro Widianto3, Ibnu Alfarobi4
1,2Computer science
1,2STMIK Nusa Mandiri
www.nusamandiri.ac.id
114002361@nusamandiri.ac.id ; 214002315@nusamandiri.ac.id ;
3,4Accounting Information Systems
3,4Bina Sarana Informatics University
www.bsi.ac.id
3kudiantoro.kdw@bsi.ac.id; 4ibnu.iba@bsi.ac.id
(*) Corresponding Author
Abstract
Customer or buyer satisfaction is closely related to how a seller of services or a store serves its visitors.
Good service for visitors also makes a good impression from visitors, while if the opposite will cause a
very bad or unfavorable impression in the eyes of customers, and may also lead to the reluctance of
visitors to come back lost the seller's opportunity to get potential buyers to become customers. This study
attempts to analyze customer satisfaction with the services provided by Arthaspa outlets in grand Kemang
hotels using the C4.5 Algorithm approach. The attributes used are comfort, cleanliness, tidiness, and price.
samples taken are customers who have transacted at least once..
Keywords: Data Mining, C4.5, Classification, Arthaspa
Abstrak
Kepuasan pelanggan atau pembeli setuju dengan layanan penjual atau penyedia layanan yang melayani
pengunjungnya. Layanan yang baik untuk pengunjung yang menarik perhatian kedua pengunjung,
sementara sebaliknya akan menarik perhatian yang lebih baik atau kurang di mata pelanggan, dan
mungkin juga menarik pengunjung yang datang kembali untuk membeli kesempatan membeli pembeli yang
berpotensi mendapatkan pelanggan. Penelitian ini mencoba menganalisis kepuasan pelanggan dengan
layanan yang disediakan oleh outlet Arthaspa yang tersedia di Grand Kemang Hotel menggunakan
Algoritma C4.5. Atribut yang digunakan adalah kenyamanan, kebersihan, kerapian, dan harga. Sampel yang
diambil adalah pelanggan yang pernah bertransaksi sekali. Studi ini mendapatkan kesimpulan paling
populer tentang simpul "harga" saat menggunakan Microsoft Excel dan "Kebersihan" saat menggunakan
perangkat lunak Rapidminer.
Kata Kunci: Data Mining, C4.5, Klasifikasi, Arthaspa.
INTRODUCTION
The best quality of service to customers,
will also affect the trust of customers to the
company so that customers are satisfied with the
services received and thus the customer will
convey satisfaction to others, with this case makes
the market share expands and the company will be
superior to its competitors (AFRIZAL, 2018).
Consumer satisfaction is a comparison
between the services received and consumer
expectations, consumers have assessed
satisfaction or dissatisfaction with the level of
expectations they create in mind. In a situation of
dissatisfaction occurs if consumers after using the
product or service purchased feel that
performance does not match consumer
expectations (Alawiyah, 2018).
Quality of service is very important and
closely related to customer satisfaction itself. With
good service quality will give satisfaction from
customers, while service quality is lacking or not
good so it gives uncomfortable effect to the
customer (Sobandi, 2019) the data visit, and it is
also possible that over time can cause customers
to switch to other competitors (Hartono,
2017)who has a similar business. The current tight
competition also forces the seller or buyer to
http://creativecommons.org/licenses/by-nc/4.0/
P-ISSN: 2656-1743 | E-ISSN: 2656-1735
DOI: https://doi.org/10.34288/jri.v2i3.137
JURNAL RISET INFORMATIKA
Vol. 2, No. 3 June 2020
102
The work is distributed under the Creative Commons Attribution-NonCommercial 4.0 International License
produce a good service so that it can satisfy the
customer (Budiarti, 2018).
Quality of service is very important and
closely related to customer satisfaction. With a
good quality of service will provide satisfaction
from customers, while a quality of service that is
lacking or not good will give an uncomfortable
effect for consumers who come to visit, and it is
also possible that over time it can cause customers
to switch to other competitors who have similar
businesses . The current competitive conditions
also force sellers or buyers to produce good
service so that they can satisfy customers.
Companies in knowledge of the market or
customers have a positive correlation in sales
performance (Noyita, 2019).
Arthaspa also prioritizes customer
satisfaction in marketing strategies for the
advancement of its company because it is related
to market share and customer retention(Puung et
al., 2014). Besides arthaspa also prioritizes good
performance by employees so that the assessment
produced by consumers of the service is also good
and it is very important for the company, because
customer satisfaction is a pleasant response and
can be fulfilled, while customer dissatisfaction is
an unexpected disappointment (Susi et al., Nd).
In this research, it is expected that the
results of customer satisfaction will be analyzed
using the data mining approach C4.5 because the
C4.5 algorithm is easy to understand and
interpreted in its use (Eki, 2016). To find out
whether Artspa has provided services that are in
accordance with customer wishes. So the
evaluation needs to be done from the customer
side considering the attributes of comfort,
cleanliness, tidiness, and price are very important
in providing satisfaction to customers.
RESEARCH METHODOLOGY
Knowledge Discovery in Database (KDD)
Data Mining or Knowledge Discovery in
Database (KDD). The KDD process is the result of
minimal data extracting a data pattern, and
altering the results so that they are easy to
understand(Riandari & Simangunsong, 2019). In
KDD there are six most basic elements in KDD
information retrieval techniques, namely:
1. Working on data that will be processed with
many sources.
2. Efficient use of data is required
3. Prioritizing statutes
4. Requires high level of language usage
5. Use several forms of automatic learning
6. Produce unique results
The Knowledge Discovery in Database
(KDD) process can be outlined as follows (Yunita,
2018) :
1. Data Selection is the process for selecting
words from a collection of data, data selection
is done before the stage of obtaining an
information in KDD. the results of the data
selection will be stored in a file, separate from
the database Pre-processing or Cleaning is
before the data mining process is carried out, it
is necessary to clean up the data that is the
focus of KDD. The cleaning process includes
removing duplicate data, checking for
inconsistent data, and correcting data errors;
2. Transformation is performed on data that has
been selected or selected. so that the data
selected is in accordance with the provisions of
the data mining process.
3. Data mining is the science of finding interesting
patterns or information in large amounts of
data by using certain techniques or methods.
techniques, methods or algorithms in data
mining vary widely. the choice of data mining
method is very dependent on the KDD process
(Mardi, 2017).
4. Interpretation or evaluation resulting from the
data mining process really needs to be defined
in a form that is easy to understand by
interested parties. this stage is one part of the
KDD process called interpretation. this stage as
an evaluation of whether the pattern or
information found is contrary to the previous
hypothesis.
Data Mining
Data mining is the science of extracting
information from big data, in accordance with the
purpose of data mining, which is to make a
decision from a large data and stored in a
database, data warehouse or information stored
from a repository (Tarigan et al., 2017).
Data collection
Data collection method is done by
observation directly and use the Questionnaire in
getting accurate data.
1. Observation is done by collecting a number of
sales data by visiting directly the hotel in
collaboration with CV. Artha Gemilang to get
the information needed related to research.
2. Questionnaire where the researcher distributes
a list of questions to respondents regarding
various aspects related to the value of each
customer satisfaction attribute which will later
be carried out on the results of research and
discussion.
http://creativecommons.org/licenses/by-nc/4.0/
JURNAL RISET INFORMATIKA
Vol. 2, No. 3 June 2020
P-ISSN: 2656-1743 |E-ISSN: 2656-1735
DOI: https://doi.org/10.34288/jri.v2i3.137
103
The work is distributed under the Creative Commons Attribution-NonCommercial 4.0 International License
Research Population
The population that will be used is the
total number of customers at outlets that work
with CV, Artha Spa since the beginning of January
2019 until the end of June 2019 who have done at
least once. Ie as many as 186 people.
Research Sample
Determination of the number of study
samples is calculated by the Slovin formula (Luvia
et al., 2017):
n =
N
1+Ne2
............................................................................. (1)
Where:
n: number of samples
N: total population
E: error tolerance limit of 5 percent or 0.05.
Number of respondents to use:
n =
200
1 + (200 ∗ 0,0025)
Yields n = 133.33, if rounded to 133. So the
number of respondents used is 127 people.
Data Analysis Method
The steps taken in data analysis include:
1 The results of the respondents' answers in the
questionnaire were converted into a Likert
scale
2 Obtained Likert Scale then to make a decision
tree with the C4.5 algorithm approach. Begin by
forming which attribute will become root, or
which root attribute will be based on the
highest gain gain. If the value of an attribute
has not resulted in a unanimous decision, then
a recalculation is made by making a new
branch under the previous node, but if a
unanimous decision has occurred, then the
calculation will be stopped and a final
conclusion obtained.
3 The results of the algorithm calculation are
then represented as a decision tree shape.
RESEARCH RESULTS AND DISCUSSION
As for the research on Arthaspa using
premiere data taken from May to June 2019 for the
results of the questionnaire can be seen in Table 1:
Table 1. Questionnaire Results
No
Attribute
Number
of cases
Yes Not
Total
50 32 18
1
Convenien
ce
Very
comfortable
31 27 4
Enough 10 5 5
Less
comfortable
9 0 9
No
Attribute
Number
of cases
Yes Not
Total
50 32 18
2
Cleanlines
s
Clean 41 32 9
Not clean 9 9 9
3 Neatness
Neat 29 23 6
Not neatly 21 9 12
4 Price
Affordable 22 22 0
Relatively
inexpensive
24 7 17
Expensive 4 3 1
Evaluation and Validation
Evaluation and validation are the results
of the classification of data that has been
determined based on the process used, henceforth
after knowing the evaluation of the classification
model based on the number of dataset records that
have been predicted correctly and incorrectly in
the classification modeling, these results can be
known as confusion matrix. after getting a number
of attributes, the next step is processing the
selection of attributes(Santoso, 2014). This
attribute selection is done to get the attributes
whose values are relevant. the following
explanation of the attributes used:
1. Convenience is an attribute given to
respondents to assess comfort in service and is
grouped into 3 categories, namely, Very
comfortable, quite, less comfortable.
2. Cleanliness is an attribute to assess the
cleanliness of the environment in the artha spa,
and is grouped into 2 categories, clean and
unclean.
3. Neatness is an attribute that assesses the
neatness of arthaspa employees who are
grouped into 2 categories, neat and untidy.
4. Price is an attribute of the price offered by
artha spas to consumers, and is categorized
into 3, namely affordable, relatively
inexpensive and expensive.
In this research the test results can be seen from
the following steps:
Information Gain and Entrhopy
The first step to do is calculate the entropy
value and information gain data in Table 2:
Table 2. Entrhopy and Gain Values for all attributes
Attribute
Amoun
t (s)
Yes
(Si)
No
(Si)
Entrho
phy
Gain
Total 50 32 18
0.9426
83
Convenience
0.398
721
Very 31 27 4 0.5547
http://creativecommons.org/licenses/by-nc/4.0/
P-ISSN: 2656-1743 | E-ISSN: 2656-1735
DOI: https://doi.org/10.34288/jri.v2i3.137
JURNAL RISET INFORMATIKA
Vol. 2, No. 3 June 2020
104
The work is distributed under the Creative Commons Attribution-NonCommercial 4.0 International License
Attribute
Amoun
t (s)
Yes
(Si)
No
(Si)
Entrho
phy
Gain
comfortable 78
Enough 10 5 5 1
Less
comfortable
9 0 9 0
Cleanliness
0.320
077
Clean 41 32 9
0.7592
76
Not clean 9 9 9 0
Neatness
.1022
92
Neat 29 23 6
0.7355
09
Not neatly 21 9 12
0.9852
28
Price
0.459
766
Affordable 22 22 0 0
Relatively
inexpensive
24 7 17
0.8708
64
Expensive 4 3 1
0.8112
78
Table 3. Price Attributes
Attribute
Amoun
t (s)
Yes
(Si)
No
(Si)
Entrho
phy
Gain
Total 50 32 18
0.9426
83
Price
0.459
766
Affordable 22 22 0 0
Relatively
inexpensive
24 7 17
0.8708
64
Expensive 4 3 1
0.8112
78
In Table 3 the Price Attributes have 3
categories, namely, affordable, relatively cheap,
expensive. With each entrapping value. because
the affordable category has a value of 0, what we
are looking for in the next node is relatively cheap
and expensive, it can be seen in Table 4:
Table 4. Relatively Cheap
Attribute
Amoun
t (s)
Yes
(Si)
No
(Si)
Entrho
phy
Gain
Total 50 32 18
0.9426
83
Convenience
0.766
425
Very
comfortable
10 7 3
0.8812
91
Enough 5 0 5 0
Less
comfortable
9 0 9 0
Cleanliness
0.626
299
Clean 16 7 9
0.9886
99
Not clean 8 0 8 0
Neatness
0.581
891
Neat 5 0 5 0
Attribute
Amoun
t (s)
Yes
(Si)
No
(Si)
Entrho
phy
Gain
Not neatly 19 7 12
0.9494
52
The results of searching for the relatively
inexpensive category price attribute yielded the
highest Gain value, namely comfort with a value of
0.766425. For further searching for the gain value
for the convenience attribute, can be seen in Table
5:
Table 5. Comfort
Attribute
Total
Qty (s) Yes (Si)
No
(si)
Entropr Gain
50 32 18 0.942683
Cleanliness
0.942683
Clean 7 7 0 0
Not clean 3 0 3 0
Neatness
0.766425
Neat 0 0 5 0
Not neatly 10 7 3 0.881291
The results of looking for comfort
attributes produced the highest Gain value, namely
cleanliness with a value of 0.942683. Because the
entropy value of the comfort category 0 for the
next calculation is to find the gain value from the
expensive category, it can be seen in Table 6:
Table 6. Expensive Categories
Total Attribute
number (s) Yes (Si)
No
(Si)
Entrapr Gain
50 32 18 0.942683
Convenience
0.877781
Very comfortable 4 3 1 0.811278
Enough 0 0 0 0
Less comfortable 0 0 0 0
Cleanliness
0.942683
Clean 3 3 0 0
Not clean 0 0 0 0
Neatness
0.877781
Neat 4 3 1 0.811278
Not neatly 0 0 0 0
Search results from the expensive
category produced the highest Gain value, namely
cleanliness with a value of 0.942683. Because the
entrhopy value of the comfort category 0 for the
node calculation is complete.
http://creativecommons.org/licenses/by-nc/4.0/
JURNAL RISET INFORMATIKA
Vol. 2, No. 3 June 2020
P-ISSN: 2656-1743 |E-ISSN: 2656-1735
DOI: https://doi.org/10.34288/jri.v2i3.137
105
The work is distributed under the Creative Commons Attribution-NonCommercial 4.0 International License
Decision Tree
Figure 1. Decision Tree Results
The results of the decision tree modeling
in Figure 1 shows that all cases are included in the
class determined and therefore the decision tree is
the final decision tree. after the results of modeling
with the decision tree the creation of rules in the
decision tree. the rules that can be formulated are
as follows:
"If Cleanliness = not clean then class = Not
satisfied"
"If Cleanliness = Clean and comfort = very
comfortable, class = satisfied"
"If Cleanliness = Clean and comfort = Less
comfortable then class = not satisfied"
"If Cleanliness = Clean and comfort = enough and
price = relatively cheap, class = not satisfied
"If Cleanliness = Clean and comfort = enough and
price = affordable, class = satisfied"
The results obtained in the sample data
that are rooted are the service attributes in the
decision tree, while attributes such as comfort,
cleanliness and price are good food. from the
sample data, the number of rules used formed 5
rules.
CONCLUSIONS AND SUGGESTIONS
Conclusion
From the results obtained in the previous
discussion, conclusions can be drawn, with the
attributes used such as: Comfort, cleanliness,
tidiness and price greatly affect the level of
customer satisfaction. If cleanliness is not clean
then the customer will be dissatisfied if the
cleanliness is clean and very comfortable then the
customer will be satisfied but if the cleanliness of
the comfort level is sufficient and the price is
affordable then the customer will be satisfied with
the existing services.
Suggestion
It needs to be evaluated and recalculated
regularly so that the company can continue to
optimize its services to customers. It is
recommended to do another algorithm approach,
included in the classification category and seen the
highest level of accuracy in order to be taken into
consideration for analyzing the next level of
customer satisfaction. And of course it can also be
a constructive suggestion
REFERENCE LIST
AFRIZAL, B. (2018). The Effect of Service Quality
and Price on Customer Satisfaction at
Warung Mbah Ganis. Master of Management,
Nusantara Business Institute, 1, 10-18.
Alawiyah, T. (2018). The Influence of Service
Quality on Consumer Satisfaction Staying at
Diamond Hotels in Samarinda City. eJournal
of Business Administration, 6 (2), 803–817.
Budiarti, A. (2018). The Effect of Service Quality
and Complaint Handling on Customer
Satisfaction and Loyalty of Sharia
Commercial Banks in Anindhyta Budiarti.
Equity, 15 (110), 210-231.
Eki, RM (2016). Customer Satisfaction Analysis
Using C4 Algorithm. 5. Faculty of Computer
Science, 3 (2), 1–14.
Hartono, A. (2017). ANALYSIS OF SERVICE
QUALITY INFLUENCE ON CUSTOMER
SATISFACTION IN FORMING CUSTOMER
LOYALTY IN THE NAKED CRAB
RESTAURANT SURABAYA. Journal of
Hospitality and Services Management, 5 (2),
174–187.
Luvia, YS, Windarto, AP, Solikhun, S., & Hartama, D.
(2017). APPLICATION OF C4.5 ALGORITHM
FOR CLASSIFICATION OF STUDENT SUCCESS
PREDICTIONS IN AMIK TUNAS NATION.
Jurasik (Journal of Information Systems and
Informatics Engineering Research), 1 (1), 75.
https://doi.org/10.30645/jurasik.v1i1.12
Mardi, Y. (2017). Data Mining: Classification Using
C4.5 Algorithm. Journal of Edict Informatics,
2 (2), 213-219.
Noyita, E. (2019). MEASURING IMPROVEMENT OF
SERVICE QUALITY AND CUSTOMER TRUST
http://creativecommons.org/licenses/by-nc/4.0/
P-ISSN: 2656-1743 | E-ISSN: 2656-1735
DOI: https://doi.org/10.34288/jri.v2i3.137
JURNAL RISET INFORMATIKA
Vol. 2, No. 3 June 2020
106
The work is distributed under the Creative Commons Attribution-NonCommercial 4.0 International License
ON CUSTOMER SATISFACTION WITH SEM
ANALYSIS METHOD (Study at Kampoeng Spa
Batam. Khazanah Ilmu Berazam, 2, 62-70.
Puung, FK, Fudholi, A., & Dharmmesta, BS (2014).
Analysis of the Effect of Service Quality on
Customer Satisfaction and Loyalty in Salons
and Spas. Analysis of the Effect of Service
Quality on Customer Satisfaction and Loyalty
in Salons and Spas, 4 (2), 105-110.
https://doi.org/10.22146/jmpf.274
Riandari, F., & Simangunsong, A. (2019).
Application of C4.5 Algorithm to Measure
Student Satisfaction Level. DIKTI Accredited,
3 (2), 1–7.
Santoso, TB (2014). Analysis and Application of
C4.5 Method For Predicting Customer
Loyalty. LIMI'S Faculty of Engineering
Scientific Journal, 10 (1), 1–6.
Sobandi, MM (2019). SERVICE MANAGEMENT IN
PHARMACIES P. BHAKTI KENCANA
TOWARDS CUSTOMER SATISFACTION.
Journal of Pharmacopolium, 2 (1).
https://doi.org/10.36465/jop.v2i1.468
Susi, I., Situmorang, S., Nasari, F., Systems, J.,
University, I., Utama, P., & Utama, UP (nd).
Analysis of customer satisfaction with spa
services with servqual (case study: pt. Royal
amadeus). December 2018, 521-528.
Tarigan, HD, Destiawati, F., & Fitriansyah, A.
(2017). Implementation of C4.5 Algorithm
Towards Customer Satisfaction. Ethos
(Journal of Research and Community
Service), Vol. 6, No, 80–86.
Yunita, F. (2018). Application of Data Mining Using
the K-Means Clustring Algorithm on
Acceptance of New Students. Systemation, 7
(3), 238.
https://doi.org/10.32520/stmsi.v7i3.388
http://creativecommons.org/licenses/by-nc/4.0/