Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

Journal of Physics: Conference Series

PAPER • OPEN ACCESS

Implementation of Data Mining to Classify the Consumer’s Complaints of


Electricity Usage Based on Consumer’s Locations Using Clustering
Method
To cite this article: A M H Pardede et al 2019 J. Phys.: Conf. Ser. 1363 012079

View the article online for updates and enhancements.

This content was downloaded from IP address 36.77.212.34 on 02/04/2021 at 16:14


The 1st Workshop on Environmental Science, Society, and Technology IOP Publishing
Journal of Physics: Conference Series 1363 (2019) 012079 doi:10.1088/1742-6596/1363/1/012079

Implementation of Data Mining to Classify the Consumer’s


Complaints of Electricity Usage Based on Consumer’s
Locations Using Clustering Method

A M H Pardede1*, Yusdiana Br Sembiring1, Akbar Iskandar2, Dyah Retno


Pitasari3, S Sriadhi4, Dian Rianita5, Muhammad Arifin6, Mersy Yoslin Ririhena7,
Nurintan Asyiah Siregar8, Supriyono9, Ayu Esteka Sari10, Simson Tondo11,
Muhammad Zarlis12, Edy Winarno13 and Tulus14
1STMIK Kaputama, Binjai, Sumatera Utara, Indonesia
2Department of Informatics, STMIK AKBA, Makassar, Indonesia
3Department of Law, Universitas Halmahera, Tobelo, Indonesia
4Department of Electrical Engineering, Universitas Negeri Medan, Indonesia
5Department of Public Administrations, Universitas Lancang Kuning, Indonesia
6Department of Information System, Faculty of Engineering, Universitas Muria Kudus,
Indonesia
7Department of Accounting, Universitas Halmahera, Indonesia
8Department of Management, STIE Labuhanbatu, Sumatera Utara, Indonesia
9Department of Information System, Universitas Muria Kudus, Kudus, Indonesia
10Department of Management, STIE Sakti Alam Kerinci, Jambi, Indonesia
11Department of Public Administration, Universitas Halmahera, Tobelo, Indonesia
12Department of Computer Science, Universitas Sumatera Utara, Medan, Indonesia
13Faculty of Information Technology, Universitas Stikubank, Semarang, Indonesia
14Department of Mathematics, Universitas Sumatera Utara, Medan, Indonesia

*akimmhp@live.com

Abstract. Data collected by PLN staff on customer’s complaints in the usage of electricity are
very huge and accumulate. Previously, there was no information about these kinds of
complaints in all sub-districts included inside the company, PT. PLN Binjai. On the other hand,
each year company experienced problems to classify all complaints in order to obtain
customers with information to later be used as a basis for making policy/decision. Data mining
used clustering method and K-Means algorithm to classify the data so that important
information is obtained from customer’s complaints on the electricity usage, using variable:
types of complaints, power using by consumer and consumer’s region. Data was analyzed
using Matlab to produce cluster centers and obtained a relationship between variables obtained
with groups with high number of complaints. Results revealed that from 500 customers who
have complained, cluster 1 had 218 complaints namely NT Fuse Putus with power used was
1300 watt and was located in Binjai.

1. Introduction
There are many kinds of complaints which are a form of protest that is given to someone’s work
because it has been done under expectation. For instance the usage of electricity by society.

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
The 1st Workshop on Environmental Science, Society, and Technology IOP Publishing
Journal of Physics: Conference Series 1363 (2019) 012079 doi:10.1088/1742-6596/1363/1/012079

Electricity users are disturbed by the problems caused by electricity that has been consumed by
consumers every single day. Every year, the complaints from consumers to PLN (Perusahaan Listrik
Negara) are increasingly fluctuating. Therefore the data obtained will vary greatly and have a large
enough data coverage and the data is in accordance with the complaints given by consumers to PLN.
With a large number of transactions, there will be data accumulation where the buildup is
underutilized by data warehouse and data mining technologies that data is needed to understand one's
activities and tendencies to plan decision-making strategies for companies by using data mining data
processing that can produce information and a useful example for PLN. One method that can be used
in the decision making process is Analytical Hierarchy Process (AHP)[1] and Hesitant Fuzzy [2]. The
decision making process can be used with website[3],[4]. The aspect that must be calculated in using
decision support system is benchmarking process[5],[6],[7] and the web security[8],[9],[10],and also
deterministic dynamic programming[11]. Decision making process can also be implemented in finding
the solution using searching algorithm[12] and to train the focus of children[13]. There are several
method that can also use in data mining. A classification process often using in data mining. The issue
that must also be handling is about class imbalance problem[14],[15]. To meet the needs of this
information system, a different system is needed from the operational database system. This is in line
with the data mining paradigm which is expected to provide fast and accurate information to support
decision making [16].
Data mining technology is one of solutions for this problem. This application utilizes the data in the
form of data accumulation of community complaints against electricity usage. From these data, it will
be processed using clustering method and the data will be analyzed using interpolation techniques.
After analyzing, patterns will be obtained in decision making. In this case, there are often difficulties
in collecting complaints from consumers to PLN to later be renewed services in accordance with the
wishes of electric power users. Therefore, this technology is needed to maximize PLN's performance
in estimating the number of community complaints with different locations so that the PLN can
monitor every complaint that exists in each region for renewal.

2. Research Methods
2.1 System Analysis
System analysis can be arranged as follows:
1. With the stages of system analysis, the system requirements can be determined so that it can
analyze the complaints collected using k-means algorithm.
2. System analysis is performed on the complaints collected using k-means algorithm.
System analysis is carried out on complaints collected using MATLAB software that has been
providing training and testing functions in the k-means algorithm [17].

2.2. Problem Analysis


The system modeling concept used by the author in designing data mining grouping for identification
process is The K-Means algorithm is the most popular and widely used clustering algorithm in the
industrial world. This algorithm is prepared on the basis of a simple idea. There are initially
determined how many clusters will be formed. The first element in the cluster can be selected as a
centroid point cluster [18]. The following is a flow chart image of the K-means algorithm. Designing
the process of consumer’s complaints. Data on electricity usage complaints will be processed by
grouping the data collected. The data needed in this analysis were a raw database results by taking
variables [19]. Non-nominal data such as data collection for answers to facilities, human resources and
convenience questions must be initialized in the form of numbers so that data can be analyzed by the
K-means method. This grouping data of consumer’s complaints can be expressed in independent
variables namely Usage Complaint (X), Power Type (Y) and Buyer Location (Z).

2
The 1st Workshop on Environmental Science, Society, and Technology IOP Publishing
Journal of Physics: Conference Series 1363 (2019) 012079 doi:10.1088/1742-6596/1363/1/012079

Table 1. Initialization of Complaint Variables

Complaints Transformation

Burning Perching 1
Damaged KWH Meter 2
Disconnected NT Fuse 3
Burning KWH 4
Trip repetition 5
Lost Contact 6
Disabled Topup 7
Burning Mcb 8
Burning Cable of JTR 9
Disconnected Incoming cables 10
Defisit 11
Disconnected Fco 12
Inverted Mcb 13
Blank KWH Meter 14
Lost contact at power pole 15
LPJ Pole ignited fire 16

Table 2. Criteria Initialization of Power

Power Transformation
450 watt 1
900 watt 2
1300 watt 3
2200 watt 4
3500 watt 5
4400 watt 6
5500 watt 7
7700 watt 8
11000 watt 9
13900 watt 10
17000 watt 11
22000 watt 12
66000 watt 13
82500 watt 14
105000 watt 15
>=164000 watt 16

3
The 1st Workshop on Environmental Science, Society, and Technology IOP Publishing
Journal of Physics: Conference Series 1363 (2019) 012079 doi:10.1088/1742-6596/1363/1/012079

Tabel 3. Initialization of Consumer's Locations

User Location Transformation

Binjai Kota 1

Binjai Timur 2

Binjai Barat 3

Non-nominal data type such as complaints, types of power and location of the user must be initialized
in the form of numbers so that the data can be processed using the K-Means method. The complaint
factor of electric power usage can be expressed in independent variables, namely Complaint (X),
Power Type (Y) and User Location (Z).

3. Results and Discussion


The data calculation of customer complaints on the use of electric power based on the location of the
power user using clustering method with this K-means algorithm, in order to generate a new
knowledge, to reveal how many groups of data of consumer’s complaints using variables kind of
complaints, the consumer’s complaints about electricity power and the areas of consumer’s. From
calculating these variables, it will reveal which areas have the most problems and provide complaints
in electricity usage. The data of customer’s complaint is needed to be inputted into Matlab
programming which is consists: data of consumer’s complaint, the complaints about electricity power
using by customer and the region of customers. The data of consumer’s complaints needed to be
collected into Microsoft Excel as a database (500 data).
The final result to determining the groups in which use is included in group1, group 2, and group 3 can
be seen in the explanation sections below:
a. Group 1 had the most complaining consumers. The results showed that the accumulation data of
Complaint (X), Power (Y), Region (Z) was 218 data. From the data description, it indicated that the
usage that often occurs is the type of use complaints with range 3, electricity power 3.63 and
Region of customers 1.40.
b. Group 2 had good result. The results of Complaint (X), Power (Y), Region (Z) was 145 data. From
the description of the data, it is shown that the usage of electricity which is often caused
consumer’s complaints with a range of 15.20, power electricity 4.23 and region of customers 1.23.
c. Group 3 also had good result. The results of Complaint (X), Power (Y), Region (Z) was 137 data.
From the description of the data, it is shown that the usage which is often caused consumer’s
complaints with a range of 9.08, power electricity 3.57 and region of customers 1.38. For more
details, it can be seen in the following table:

Table 4. Results (Cluster 3)

NO X Y Z Group NO X Y Z Group
1 10 2 3 3 22 7 7 3 3
2 8 4 1 3 23 10 2 1 3
3 8 4 1 3 24 8 2 1 3
4 8 2 1 3 25 8 8 1 3
5 10 2 1 3 26 8 2 1 3
6 7 2 1 3 27 10 2 1 3
7 9 2 2 3 28 9 2 1 3
8 7 4 1 3 29 9 2 1 3
9 9 4 3 3 30 8 2 1 3

4
The 1st Workshop on Environmental Science, Society, and Technology IOP Publishing
Journal of Physics: Conference Series 1363 (2019) 012079 doi:10.1088/1742-6596/1363/1/012079

10 9 2 2 3 31 10 2 1 3
11 8 4 1 3 32 11 2 1 3
12 10 4 1 3 33 10 2 1 3
13 11 4 1 3 34 8 5 1 3
14 10 4 1 3 35 10 4 1 3
15 7 7 1 3 36 7 1 1 3
16 8 2 2 3 37 9 2 1 3
17 9 2 2 3 38 9 2 1 3
18 10 2 2 3 39 9 2 1 3
19 11 7 1 3 40 9 7 1 3
20 7 7 1 3 41 10 2 1 3
21 7 2 3 3 42 8 2 1 3
NO X Y Z Group NO X Y Z Group
43 10 2 1 3 86 11 2 3 3
44 9 2 1 3 87 11 7 3 3
45 10 8 1 3 88 11 2 3 3
46 9 2 1 3 89 7 7 3 3
47 10 2 1 3 90 10 7 3 3
48 10 8 1 3 91 8 2 1 3
49 9 10 1 3 92 8 7 3 3
50 10 4 1 3 93 8 2 1 3
51 8 2 1 3 94 10 2 1 3
52 10 2 1 3 95 7 2 1 3
53 8 2 1 3 96 9 2 1 3
54 10 2 1 3 97 7 2 1 3
55 9 2 1 3 98 9 7 1 3
56 9 7 1 3 99 9 2 1 3
57 10 7 1 3 100 8 2 1 3
58 9 5 1 3 101 9 7 1 3
59 9 8 1 3 102 8 2 1 3
60 9 2 1 3 103 10 2 1 3
61 10 2 1 3 104 11 7 1 3
62 8 2 1 3 105 8 10 1 3
63 10 2 1 3 106 8 2 2 3
64 8 5 1 3 107 8 2 2 3
65 8 2 1 3 108 8 7 1 3
66 8 2 1 3 109 9 3 2 3
67 10 2 1 3 110 11 7 3 3
68 9 7 1 3 111 11 2 3 3
69 9 2 1 3 112 7 7 3 3
70 9 2 1 3 113 8 2 3 3
71 10 8 1 3 114 9 2 2 3
72 8 2 1 3 115 9 2 1 3
73 9 2 1 3 116 9 3 2 3
74 7 3 1 3 117 11 2 3 3
75 11 3 1 3 118 11 7 3 3
76 11 3 1 3 119 11 2 3 3
77 8 3 1 3 120 7 7 3 3
78 8 2 1 3 121 7 7 1 3
79 10 2 1 3 122 11 2 1 3
80 11 2 2 3 123 11 2 1 3

5
The 1st Workshop on Environmental Science, Society, and Technology IOP Publishing
Journal of Physics: Conference Series 1363 (2019) 012079 doi:10.1088/1742-6596/1363/1/012079

81 11 7 2 3 124 8 8 1 3
82 11 2 2 3 125 8 2 2 3
83 7 7 1 3 126 10 2 1 3
84 8 2 2 3 127 11 2 1 3
85 10 2 1 3 128 11 2 1 3
NO X Y Z Group NO X Y Z Group
129 11 7 1 3 134 11 7 1 3
130 7 2 1 3 135 11 2 1 3
131 8 2 1 3 136 7 2 1 3
132 10 2 1 3 137 9 2 1 3
133 11 7 3 3

4. Conclusion
From the results, it can be identified that Cluster 1 had the highest number of complaints, 269 users.
The type of complaint of from these costumers was Type of NT Fuse Disconnect with the type of
power used was 1300 watt, and it was located in the District of Binjai City. Also, the group which had
the highest numbers of complaining costumers who experienced problems in the use of electricity was
Cluster 1 with 218 data users who often experience complaints in the group Type of Complaints are
Type of NT Fuse Disconnect Complaints with the type of power that is used was 1300 watt, and was
located in Blam or Binjai City.

References
[1] A. Alesyanti, R. Ramlan, H. Hartono, and R. Rahim, “Ethical decision support system based on
hermeneutic view focus on social justice,” International Journal of Engineering & Technology,
vol. 7, no. 2.9, pp. 74–77, 2018.
[2] T. Simanihuruk et al., “Hesitant Fuzzy Linguistic Term Sets with Fuzzy Grid Partition in
Determining the Best Lecturer,” International Journal of Engineering & Technology, vol. 7, no.
2.3, pp. 59–62, Mar. 2018.
[3] R. Sitompul, A. Alesyanti, H. Hartono, and A. S. Ahmar, “Revitalization Model The Role of Tigo
Tungku Sajarangan in Fostering Character of Children in Minangkabau Family and Its
Socialization Through Website,” International Journal of Engineering & Technology, vol. 7, no.
2.5, pp. 53–57, Mar. 2018.
[4] R. Sitompul, A. Alesyanti, H. Hartono, and R. Rahim, “Legal Protection for Children Born from
Unregistered Marriage in Medan City and Its Socialization Through Website,” International
Journal of Engineering & Technology, vol. 7, no. 2.14, pp. 246–250, 2018.
[5] D. Abdullah, Tulus, S. Suwilo, S. Effendi, and Hartono, “DEA Optimization with Neural Network
in Benchmarking Process,” IOP Conference Series: Materials Science and Engineering, vol.
288, p. 012041, Jan. 2018.
[6] D. Abdullah, T. Tulus, S. Suwilo, S. Efendi, M. Zarlis, and H. Mawengkang, “A Research
Framework for Data Envelopment Analysis with Upper Bound on Output to Measure Efficiency
Performance of Higher Learning Institution in Aceh Province,” International Journal on
Advanced Science, Engineering and Information Technology, vol. 8, no. 2, 2018.
[7] D. Abdullah, Tulus, S. Suwilo, S. Efendi, Hartono, and C. I. Erliana, “A Slack-Based Measures
for Improving the Efficiency Performance of Departments in Universitas Malikussaleh,”
International Journal of Engineering & Technology, vol. 7, no. 2, pp. 491–494, Apr. 2018.
[8] R. Rahim et al., “Combination Base64 Algorithm and EOF Technique for Steganography,” J.
Phys.: Conf. Ser., vol. 1007, no. 1, p. 012003, 2018.
[9] M. Mesran et al., “Combination Base64 and Hashing Variable Length for Securing Data,” J.
Phys.: Conf. Ser., vol. 1028, no. 1, p. 012056, 2018.

6
The 1st Workshop on Environmental Science, Society, and Technology IOP Publishing
Journal of Physics: Conference Series 1363 (2019) 012079 doi:10.1088/1742-6596/1363/1/012079

[10] D. Abdullah et al., “Super-Encryption Cryptography with IDEA and WAKE Algorithm,” J. Phys.:
Conf. Ser., vol. 1019, no. 1, p. 012039, 2018.
[11] D. Abdullah, R. Rahim, D. Hartama, A. Abdisyah, Z. Zulmiardi, and S. Efendi, “Application of
Web Based Book Calculation using Deterministic Dynamic Programming Algorithm,” J. Phys.:
Conf. Ser., vol. 1019, no. 1, p. 012040, 2018.
[12] R. Rahim et al., “Breadth First Search Approach for Shortest Path Solution in Cartesian Area,” J.
Phys.: Conf. Ser., vol. 1019, no. 1, p. 012036, 2018.
[13] M. F. Syahputra et al., “Implementation of augmented reality to train focus on children with s
pecial needs,” J. Phys.: Conf. Ser., vol. 978, no. 1, p. 012109, 2018.
[14] Hartono, O. S. Sitompul, Tulus, and E. B. Nababan, “Optimization Model of K-Means Clustering
Using Artificial Neural Networks to Handle Class Imbalance Problem,” IOP Conference Series:
Materials Science and Engineering, vol. 288, p. 012075, Jan. 2018.
[15] Hartono, O. S. Sitompul, E. B. Nababan, Tulus, D. Abdullah, and A. S. Ahmar, “A New Diversity
Technique for Imbalance Learning Ensembles,” International Journal of Engineering &
Technology, vol. 7, no. 2, pp. 478–483, Apr. 2018.
[16] Eko Prasetyo. 2012. Data Mining : Konsep Dan Aplikasi Menggunakan Matlab, Yogyakarta,
Penerbit CV Andi
[17] Fajar Astuti Hermawati, Data Mining, Penerbit CV Andi, Yogyakarta 2013
[18] Jong jeng siang. 2009, Data Mining Terapan dengan Matlab, graha ilmu, Yogyakarta 2007
[19] Fajar Astuti Hermawati, Data Mining, Penerbit CV Andi, Yogyakarta 2013

You might also like