JURNAL INFOTEL

Informatics - Telecommunication - Electronics

Website: https://ejournal.ittelkom-pwt.ac.id/index.php/infotel
ISSN: 2085-3688; e-ISSN: 2460-0997

Implementation of association rule using apriori algorithm

and frequent pattern growth for inventory control
Imam Riadi1 , Herman2 , Fitriah3,* , Suprihatin4 , Alwas Muis5 , Muhajir Yunus6
1,4
Department of Information System, Universitas Ahmad Dahlan
2,3,5,6
Master Program of Informatics, Universitas Ahmad Dahlan
1,2,3,4,5,6
Jl. Ahmad Yani, Tamanan Banguntapan, Bantul 55191, Indonesia
*
Corresponding email: [email protected]

Received 8 June 2023, Revised 13 November 2023, Accepted 21 November 2023

Abstract — Inventory control is needed to maintain smooth operations and improve product efficiency. The problem of
procurement of goods that are not by needs is a problem that often occurs in various companies. This problem has the
potential to affect business processes that hinder increased sales. Based on this, a strategy is needed in managing inventory
to increase customer satisfaction. This study aims to find combinations of goods to be analyzed, resulting in an Association
Rule that can be used by store owners to manage the inventory of goods. The algorithms used in this study are Apriori and
Frequent Pattern Growth (FP-Growth). Apriori algorithm methods and FP-Growth are two techniques used in data mining to
find combinations of goods resulting in the Association Rule. This research begins with literature study and data acquisition
as the first step in data pre-processing. Next, data analysis is performed to identify patterns and relationships between items
in the dataset. At the evaluation stage, the lift ratio method is used to determine the validity of association rules. This is
an important step to ensure that the rules found have relevance and are being analyzed so that they can be used for better
decision-making. The results of the implementation of a priori and FP-Growth algorithms for 2-item analysis show that they
both produce the same number of association rules, i. e. 16 rules. There is a similarity in the rules with the highest lift ratio,
which is 14. In addition, some other rules with a lift ratio of 5 are also found in both algorithms. Next, the calculation of
the 3-item set lift value, both show the same lift ratio value. This indicates that they are effective in extracting association
rules from the same dataset, for both 2-itemset and 3-itemset.
Keywords – apriori, association rule, data mining, frequent pattern growth, lift ratio

Copyright ©2023 JURNAL INFOTEL

All rights reserved.

I. I NTRODUCTION tion technology has begun to be utilized by competition

between rulers, so developers need to find the right
Data is very important in today’s era and can strategy to meet needs and ensure customer satisfac-
support decisions on various aspects. Making decisions tion. One of the strategic innovations in business is
based on facts from data in the field is a smart move to look for connections or linkages between different
and can improve the quality of decisions. One example sets of products so that they can be packaged and
is in the sales process, especially in retail stores that sold simultaneously [5]–[7]. This innovation provides
provide various products [1], [2]. Every day, retail a solution to sales and inventory problems because
stores make many sales transactions, so the sales data product mismatches increase sales. However, finding
generated is also very much. However, the data should relationships between goods is a difficult process due
not only be used as a record without added value but to large product problems, especially when merchants
must be used for business purposes. Transaction data have thousands of products [8]–[11].
storage will continue to grow over time, so data must
be optimized to improve efficiency [3], [4]. For inventory management, data mining algorithms
can be used to find patterns in inventory data, one of
Goods control is an important aspect that needs which is by applying the association rule technique and
to be managed properly so that a business can run the method that can be applied in this study is market
smoothly and efficiently. The development of informa- basket analysis [12]–[14]. Market basket analysis is

Jurnal Infotel, Vol. 15, No. 4, November 2023

369
https://doi.org/10.20895/infotel.v15i4.980
 ISSN: 2085-3688; e-ISSN: 2460-0997
Implementation of association rule using apriori algorithm and frequent pattern growth for inventory control

Fig. 1. Research flow.

one method in data mining [15], [16]. Data mining is a rithms, especially when the dataset is very large and
complex data analysis process that involves statistical, complex. These two algorithms have the same goal of
mathematical, and artificial intelligence techniques to finding patterns in datasets, but apriori and FP-Growth
explore and discover hidden patterns in large data algorithms have different approaches and techniques
sets [17]–[19]. The goal of data mining is to get valu- for extracting relevant information from data. Natural
able and useful information from data to make business inventory control [37]–[40].
decisions. Data mining can be applied to different
Some previous research has been related to the
types of data such as customer data, inventory data,
utilization of association rule algorithms. For example,
transaction data, marketing data, and other data [20]–
research conducted by Ardiantoro and Sunarmi [41]
[22].
(2019) aims to analyze the game patterns of badminton
The data mining process includes several steps players which is one of the popular sports in Indone-
such as data preprocessing, selection of appropriate sia. In this study, data was collected and processed
data mining methods, method execution, evaluation of by dividing the playing field into several play areas.
results, and interpretation of results. Some commonly In addition, Apridonal et al. [42] (2019) conducted
used data mining techniques include classification, research to develop an application that can identify
regression, clustering, association, and ranking models. sales patterns using the Association Rule method and
Each of these techniques has a different purpose and apriori algorithm. Hu et al. [43] also used frequency
can be applied to different types of data [23], [24]. pattern growth algorithms in their study to analyze the
In business, data mining can help optimize inventory stability of public transport trips using association rule
management, improve product efficiency, improve cus- mining. The next research was conducted by Elisa and
tomer experience, and make more accurate business Azwanti [44] (2019) who used the association rule
decisions [25], [26]. Data mining can also help identify method with the FP-Growth algorithm to analyze the
hidden trends and patterns in data, allowing companies frequency of purchasing 3 kg LPG gas.
to respond more effectively to changing market or Based on previous research that has used FP-Growth
customer needs [27]–[32]. and apriori algorithms for various purposes, such as an-
Two data mining algorithms that are often used alyzing game patterns, travel stability, developing chil-
in inventory analysis are the apriori algorithm and dren’s toy applications, purchasing fertilizers, trans-
FP-Growth. Both of these algorithms can be used portation trip stability, and purchasing 3 kg LPG gas,
in the context of inventory control to find patterns the FP-Growth algorithm is more often used because
of association and high frequency between inventory it is considered superior to apriori [45]. Therefore, this
items [33]–[35]. Apriori algorithm is an algorithm in study will also use the FP-Growth algorithm to find
data mining that is used to find patterns that often association rules to help store owners manage their
appear (frequent itemsets) in a collection of transaction product inventory. Although using the same algorithm,
data. This algorithm works by dividing the dataset the main difference lies in the use of different data,
into smaller subsets and then looking for items that namely clothing sales data from the city of Bengkulu.
often appear in each subset. After that, the found The purpose of this study is to find combinations of
itemsets will be combined to form larger itemsets and goods so that store owners know the right placement of
then recalculated in frequency. This process is done goods and analyze itemsets to obtain association rules
iteratively until no more frequently occurring itemsets using apriori and FP-Growth algorithms.
can be found. Apriori algorithms are useful in data
analysis and marketing, where we can find frequently II. R ESEARCH M ETHOD
emerging consumer buying patterns and create more This research stage begins with a literature review
effective marketing strategies [36]. and then data acquisition is carried out for processing.
Next, analyze by implementing the apriori algorithm
The FP-Growth algorithm uses a top-down approach and FP-Growth in goods control. The research flow
to generate a tree data structure called FP-Tree, which can be seen in Fig. 1.
represents all transactions in a more concise form.
Then, the FP-Growth algorithm will extract frequent A. Literature Study
itemsets from the FP-Tree using a recursive approach, The study of literature is a relatively simple research
which is a more efficient alternative to apriori algo- method, but it is very important in this process of

Jurnal Infotel, Vol. 15, No. 4, November 2023

370
https://doi.org/10.20895/infotel.v15i4.980
 ISSN: 2085-3688; e-ISSN: 2460-0997
Implementation of association rule using apriori algorithm and frequent pattern growth for inventory control

Table 1. The Name of the Item in the GalleryNCA Store

inquiry. This method involves collecting data from Initials Item name Initials Item name
various literature sources, such as journals, books, K Kaos BBH Baby hanger
papers, theses, theses, and related research articles. RI Rok include KKB Kaos kaki baby
KL Kaos lemon KJP Kemeja jeans pdk
Data collection is done by reading, recording, and pro- LS Lejing s. M. L CA Celana amrico
cessing information from these sources. The process of SC Stelan clobie SHC Stelan carnaval
data analysis in literature studies begins with evaluating SCP Stelan clonie p LX Lejing xl
CAK Cd agree kids LB1 Lejing baby 123
the most relevant and current scientific work with this
DK Daster kalong ST0 Bh2telan timy 0
research. The selection of literature starts from the CKN Celana katun KS Kemeja salur
latest and passes on to the older ones. Any previous CK Celana kerja HY Hayami
research relevant to this topic will be summarized to RB Rok baloteli KD Kaos destino
BKN Blus katun STS Stelan timy s.l
evaluate the alignment of the issues raised with the BF Blus filimili SCO Shot codigo
issues studied in this study. An important and relevant HL Handuk lancer BH Bh kawat
part of the research problem is also considered. By PS Pp sielie/meong CVM Cd vaya m.l
referring to the relevant literature, this research can KN Kaos noren BI Blus impor
STX Stelan timy xl DKS Dres kaos
design a system that can be used for inventory control LK Lejing korea TI Tunik impor
based on the association rule method with apriori CKJ Celana kerja J GD7 Gamis de 7
algorithm and FP-Growth. JV Jamper valvet CKR Cp karakter
B Blus TK Tunik katun
BK Blus korea SVS Stelan velvet s.m.
B. Data Acquisition RKN Rok katun SS Stelan sibulat
TP Tunik platinum J Jumper
Data acquisition, the process of collecting data from EJ Egoist jumbo CT Cp tessa
various sources, holds significant benefits in various BS Baby shop DHK Daster hap katun
SI Stelan impor H Hoki
contexts. Firstly, data acquisition provides essential SM Stelan mld GI Gaun impor
information needed for making informed decisions, be SMI Stelan milady
it in business, research, or scientific fields. Further-
more, the data gathered can be used for monitoring
the performance of systems, devices, or processes, as and evaluate sales performance. The following item
well as for analysis and pattern discovery that can names in the GalleryNCA store are shown in Table 1.
help identify trends, weaknesses, opportunities, and
threats. Data acquisition also plays a crucial role in C. Preprocessing
scientific research, where the obtained data is used Before data can be processed with apriori and
to support or refute hypotheses, identify trends, and FP-Growth algorithms, a crucial first step is to pre-
draw scientific conclusions. In business, data acquired process the data. Pre-processing includes two main
through acquisition can be used to optimize processes, stages, namely cleaning and data transformation. The
reduce costs, and enhance quality. It also supports cleaning phase includes deleting invalid data, grouping
better decision-making, reducing uncertainty in the missing values, or handling outlier data to ensure good
decision-making process. Additionally, data acquisition data quality. In addition, the data transformation stage
fuels the development of intelligent systems such as includes changing the data format, normalization, or
artificial intelligence (AI) and machine learning, which encoding categorical data to numerical form so that
are used for classification, prediction, and pattern it can be used by the algorithm. By careful pre-
recognition. The collected data is also employed for processing, the data will be more readily used by
security purposes, detecting anomalies, and developing apriori and FP-Growth algorithms, resulting in more
products that better match customer needs. Data acqui- accurate and informative analysis results. Thus, pre-
sition also has a role in compliance and audit, used to processing data becomes an important step in the
track and verify adherence to applicable regulations process of extracting association patterns and analyzing
and standards. Thus, data acquisition is a key step in itemsets.
optimizing various aspects of life and business.
D. Analysis
The clothing data acquisition process begins with
the request submission stage which aims to obtain a This study conducts process analysis using data
complete copy of sales data from the cash register mining methodology by applying apriori and FP-
system. This data includes all clothing sales transac- Growth algorithms to determine the frequent itemset
tions recorded in the period from February 1 to March that will be used as a basis in formulating the best
31, 2023. During that period, this data will include rules or recommendations in purchasing goods.
detailed information about each transaction, including 1) Stages of implementation of apriori algorithm
details about products sold, number of items recorded, The first step in forming association rules is to
and total transactions. This data will be the basis for determine the minimum value of support to get high
a more in-depth analysis, allowing the owner of the frequency. The higher the support value, the more often
figure to understand sales trends, manage inventory, the itemset appears, and this can be useful information

Jurnal Infotel, Vol. 15, No. 4, November 2023

371
https://doi.org/10.20895/infotel.v15i4.980
 ISSN: 2085-3688; e-ISSN: 2460-0997
Implementation of association rule using apriori algorithm and frequent pattern growth for inventory control

in association analysis, in which case the researcher • Conditional FP-Tree development phase: FP-
sets the minimum support value at 0.2 % with (1). Growth builds a data structure known as a con-
P
Contains A ditional FP-Tree. This structure describes the
Support(A) = P × 100% (1) relationship between itemsets that appear together
T otal T ransactions
in a transaction.
The formula is used to measure the extent to which • Frequent itemset search phase by utilizing TID
the itemset or element (A) appears in a dataset transac- (Transaction Identity): In this stage, FP-Growth
tion. Data that has met the support set of frequently oc- uses the conditional FP-Tree that has been built
curring items is merged, and the set of non-conforming to efficiently generate frequent itemset. This in-
items will be eliminated, and the set that meets will cludes the calculation of support and confidence
be used in iterations. The second step is to search values.
for 2-itemset. The value of the 2-itemset is calculated E. Evaluation
using (2).
P Association analysis is the validity of a rule de-
Contains A and B termined using the lift ratio method. Lift ratio is
Support (A → B) = P ×100%
T otal T ransaction a common indicator used to determine whether an
(2) association rule is considered valid or not. An associ-
The value containing A and B is the value of the ation rule is considered valid when the lift ratio value
two items of goods in the transaction, dividing the in the rule exceeds 1, and the higher the lift ratio
value containing A and B by the total amount of all value, the stronger the rule. In apriori and FP-Growth
transactions. The third step 3-itemset support value is algorithms, lift ratio is used to evaluate the strength
obtained using (3). of the association rule found to calculate lift ratio can
use (5).
Support (A,B, C) =
P Conf idence(A, B)
Contains A, B, and C Lif t Ratio =
= P × 100% Benchmark Conf idence(A, B)
T otal T ransaction (5)
(3)
For benchmark confidence use (6).
Confidence is one of the important aspects of asso- Nc
ciation analysis, which is used to measure the extent to Benchmark Conf idence = (6)
N
which an association or association rule applies. In this
context, we will understand how strong the relationship III. R ESULT
between two itemsets, namely A and B or A, B and
This section discusses research data, pre-processing,
C. Confidence, measured in percentages, gives an idea
apriori implementation, and FP-Growth implementa-
of how often itemset B appears in transactions that
tion.
also contain itemset A or how often Item C appears in
transactions that also contain items A and B, in this A. Research Data
study the minimum value of confidence has been set Data collection in this study was carried out through
at (0.8 %) To determine the value of confidence, it is clothing sales transactions recorded in the form of
obtained using (4). shopping receipts. Every time a customer makes a pur-
Conf idence = P (A | B) = chase of clothes in the store, the information recorded
N umber of T ransactions A and B in the shopping receipt includes details such as the
= × 100% type of clothes purchased, the number of items, the
N umber of T ransactions A
(4) transaction number. The data used was 140 transactions
and 57 items of goods which were sales samples at the
2) Stages of FP-Growth algorithm implementation GalerryNca store in Bengkulu City. Transaction dataset
The search for associations using the FP-Growth in tabular form as in Table 2.
method is a development of apriori algorithms. In the
FP-Growth algorithm, confidence and support values Table 2 is a dataset of goods transactions, the name
are applied similarly to apriori, but the difference lies of the goods has been changed to initials. To find out
in the more efficient approach used by FP-Growth. the full name of each item, refer to Table 1.
FP-Growth reduces complexity by involving only one B. Pre-processing
iteration through a dataset of transactions, in contrast to
apriori which requires multiple iterations. FP-Growth The data processing process is carried out with the
implements three main stages: aim of finding patterns of similarity of purchased goods
based on daily purchase report data. Sales report data
• Conditional pattern base generation phase: At this every day at first is just ordinary information then
stage, FP-Growth identifies patterns that appear after processing it will be very useful information for
repeatedly in the transaction dataset. future business progress and improvement. Therefore,

Jurnal Infotel, Vol. 15, No. 4, November 2023

372
https://doi.org/10.20895/infotel.v15i4.980
 ISSN: 2085-3688; e-ISSN: 2460-0997
Implementation of association rule using apriori algorithm and frequent pattern growth for inventory control

Table 2. Transaction Dataset Table 4. Minimum Support 1-itemset

Transaction Name of Goods Transaction Items Support (%)
No Item 1 Item 2 Item 3 Item 4 LK 0.179 17.86
210105.001 K - - - CK 0.121 12.14
210105.002 RI K - - STX 0.114 11.43
210105.003 KL - - - RI 0.086 8.57
210105.004 LS SC SCP - BF 0.071 7.14
210105.005 CAK - - - CAK 0.071 7.14
··· ··· ··· ··· ··· HL 0.071 7.14
210205.009 GI ST KJK CA BK 0.064 6.43
210208.010 K BS - - K 0.057 5.71
210208.011 BKN BF - - B 0.043 4.29
210208.012 HL RKN - - PS 0.043 4.29
210208.013 PS CK TP - JV 0.036 3.57
210208.014 KN PS RB - KN 0.036 3.57
210208.015 EJ K SI - EJ 0.029 2.86
LS 0.029 2.86
Table 3. Product Transaction Tabulation Results RB 0.028 2.86
No. Transaction Good Items
BKN 0.021 2.14
No K RI ··· G
CKJ 0.021 2.14
1. 210.105.001 1 0 ··· 0
DK 0.021 2.14
2. 210.105.002 1 1 ··· 0
LB1 0.021 2.14
3. 210.105.003 0 0 ··· 0
LX 0.021 2.14
4. 210.105.004 0 0 ··· 0
RKN 0.021 2.14
5. 210.105.005 0 0 ··· 0
SCP 0.021 2.14
6. 210.105.006 0 0 ··· 0
SI 0.021 2.14
··· ··· ··· ··· ··· ···
SMI 0.021 2.14
140. 210.208.015 0 0 ··· 0
ST0 0.021 2.14
TP 0.021 2.14

researchers will test the results of using apriori and FP-

Growth to determine the relationship between buying Table 5. Minimum Support 2-itemset
Transaction Items Support (%)
patterns. LK STX 0.1 10
LK RI 0.071 7.1
In the sales data table, the product sold has been LK CAK 0.057 5.7
processed into a tabular form with binary contents, a STX RI 0.036 3.6
value of 1 indicates that the product has been sold, STX CAK 0.021 2.1
while a value of 0 indicates that the product has not BF BK 0.021 2.1
BF BKN 0.021 2.1
been sold. The data view can be seen in Table 3. From CK RB 0.021 2.1
a tabular format it is easier to know how many items
were purchased in each transaction to look for high-
frequency patterns by creating a 1-itemset.
with (3) display the results of the 3-itemset calculation
C. Apriori Implementation
in Table 6. From the results of the calculation of the
1) High frequency minimum Support 3-itemset, then the determination of
The stages of finding a high frequency with a association rules will be carried out from 2-itemset and
predefined minimum support (0.2 %) using (1), the 3-itemset data that meet the minimum value of support
number of transaction items (A) divided by the total and minimum confidence.
number of transactions and multiplied (100 %), items 4) Apriori association rules
that do not meet the minimum support will be removed The association rule is obtained by calculating us-
and items that meet the minimum support can be seen ing (4) based on the results of the 2-itemset value that
in Table 4. satisfies the support and confidence values, the results
2) Minimun support 2-itemset can be seen in Table 7.
Table 4 displays items that have met the minimum
support for 1-itemset then a calculation will be carried
out with 2-itemset where items are purchased simulta- A total of 16 2-itemset association rules are obtained
neously in one transaction as many as 2 items (A and from the minimum support and minimum confidence
B) with the same minimum support provisions (0.2 %). values that have been determined. Furthermore, the
The calculation results using (2) are shown in Table 5. results of the association rule with the 3-itemset are
After getting the result of the 2-itemset then proceed shown in Table 8. Based on the results of the analysis,
to look for a combination of 3-itemset. 24 associations rules 3-itemset were obtained.
3) Minimum Support 3-itemset
Results Transaction data with 2-itemset shown in
Table 6. Minimum Support 3-itemset
Table 4 that has met the minimum support value Transaction Items Support (%)
will be continued with the calculation of 3-itemset, LK STX RI 0.036 3.60 %
by calculating transactions of items A, B, and C LK STX CAK 0.021 2.10 %

Jurnal Infotel, Vol. 15, No. 4, November 2023

373
https://doi.org/10.20895/infotel.v15i4.980
 ISSN: 2085-3688; e-ISSN: 2460-0997
Implementation of association rule using apriori algorithm and frequent pattern growth for inventory control

Fig. 2. FP-Tree formation results.

Table 7. Association Rules Apriori 2-itemset D. FP-Growth Implementation

No. Rules Support Confidence
(%) (%) 1) Creation of FP-Tree trees
1 If buying LK → STX 10 56 From data that has met the minimum value of
2 If buying STX → LK 10 88
3 If buying LK → RI 7 40 high-frequency support. Then the items are sorted by
4 If buying RI → LK 7 83 priority as seen from the largest frequency value, sorted
5 If buying LK → CAK 6 32 according to priority as shown in Table 9. Next, the
6 If buying CAK → LK 6 80 data that has been sorted by priority will be formed into
7 If buying STX → RI 4 31
8 If buying RI → STX 4 42 an FP-Tree tree. It starts from Null and will be followed
9 If buying STX → CAK 2 19 by an itemset in the transaction until all transactions
10 If buying CAK → STX 2 30 form a tree as shown in Fig. 2.
11 If buying BF → BK 2 33
12 If buying BK → BF 2 33 Table 9. Item Transactions by Priority
13 If buying BF → BKN 2 30 No. Transaction Code Items
14 If buying BKN → BF 2 30 1 210105.001 K
15 If buying CK → RB 2 18 2 210105.002 RI, K
16 If buying RB → CK 2 75 3 210105.004 LS, SCP
4 210105.005 CAK
5 210105.006 CK, DK
Table 8. Association Rules Apriori 3-itemset 6 210105.007 CK, RB
No. Rules Supp- Confiden- 7 210105.008 BF, BK
ort (%) ce (%) 8 210105.009 HL
1 If buying LK → STXˆRI 2 20 9 210105.010 PS
2 If buying LK → RIˆSTX 2 20 10 210105.011 KN
3 If buying LKˆSTX → RI 2 20 ··· ··· ···
4 If buying LKˆRI → STX 2 20 122 210208.013 CK, PS, TP
5 If buying STX → LKˆRI 2 31 123 210208.014 PS, KN, RB
6 If buying STX → RIˆLK 2 31 124 210208.015 K, EJ, SI
7 If buying STXˆLK → RI 2 31
8 If buying STXˆRI → LK 2 31 2) Generating conditional pattern base
9 If buying RI → LKˆSTX 2 42
10 If buying RI → STXˆLK 2 42
At this stage, the FP-Growth algorithm will break
11 If buying RIˆSTX → LK 2 42 down the FP-Tree results in Fig. 2 based on each suffix
12 If buying RIˆLK → STX 2 42 and produce as mentioned in Table 9, and the item
13 If buying LK → STXˆCAK 4 12 names will be used as code for the next step sorted
14 If buying LK → CAKˆSTX 4 12
15 If buying LKˆCAK → STX 4 12 from small to largest.
16 If buying LKˆSTX → CAK 4 12 3) Generating conditional FP-Tree
17 If buying STX → LKˆCAK 4 19 To find the conditional FP-Tree, the step is to sum
18 If buying STX → CAKˆLK 4 19
19 If buying STXˆCAK → LK 4 19
the number of existing supports, and the larger number
20 If buying STXˆLK → CAK 4 19 of supports will replace them with the conditional FP-
21 If buying CAK → LKˆSTX 4 30 Tree. The support count is the number of occurrences
22 If buying CAK → STXˆLK 4 30 of an itemset in all transactions in the dataset. Itemsets
23 If buying CAKˆLK → STX 4 30
24 If buying CAKˆSTX → LK 4 30 with a larger number of supports are considered more
important because they appear more frequently in the

Jurnal Infotel, Vol. 15, No. 4, November 2023

374
https://doi.org/10.20895/infotel.v15i4.980
 ISSN: 2085-3688; e-ISSN: 2460-0997
Implementation of association rule using apriori algorithm and frequent pattern growth for inventory control

Table 10. Conditional Pattern Base Table 12. Frequent Itemset

Item Conditional Pattern Base
CKJ {{LK, STX, CAK :2} {:1}} Item Conditional Conditional Frequent
BKN {{BF:3}} Pattern Base FP-Tree Itemset
SCP {{CK:1}, {LS:2}} BKN {BF:3} BF:3 {BF-BKN: 3}
SI {{K, EJ:1}, {:2}} RB {{PS, KN:1}, CK:3 {CK-RB:3}
SMI {{SI:1}, {:2}} {CK:3}}
LX {{:3}} BK {{RI:1}, {BF:3}, BF:3 {BF-BK:3}
LB1 {{LS:2}, {:1}} {:5}}
ST0 {{K:2}, {:1}} CAK {{:2}, {LK, CK, LK:5, {LK-CAK:8,
RKN {{B, JV:1}, {:2}} STX:1}, {LK, STX:3 STX-CAK:3}
TP {{:3}} STX:2}, {LK:5}} {LK, STX,
DK {{CK:2}, {:1}} CAK:3}
EJ {{K:1}, {:3}} RI {{LK, STX:5}, LK-STX:3, {LK, STX, RI:5}
LS {{:4}} {LK:7}, {CK:1}, LK: 7 {LK, RI :10}
RB {{PS, KN:1}, {CK:3}} {:3}
JV {{B, JV:2}, {:4}} STX {{LK: 14}, {LK, LK:14 {LK-STX:14}
CK:1}, {:2}}
KN {{PS:1}, {RI:1}, {:3}}
PS {{CK:1}, {RI, HL:1}, {:4}}
B {{LK, RI:1}, {:5}} Table 13. Support 2-itemset
K {{CK, RI:1}, {RI:1}, {:6}} Items Transaction Total Support
BK {{RI:1}, {BF:3}, {:5}} Transactions (%)
CAK {{:2}, {LK, CK, STX:1}, {LK, STX:1}, {LK:7}} BF-BKN 3 140 2
BF {{:10}} CK-RB 3 140 2
HL {{:8}, {RI:1}, {LK:1}} BF-BK 3 140 2
RI {{LK, STX:5}, {LK :7}, {CK:1}, {:3}} LK-CAK 8 140 6
STX {{LK: 14}, {LK, CK :1}, {:1}} STX-CAK 3 140 2
CK {{LK:1}, {:16}} LK-RI 7 140 5
LK {{:25}} STX-RI 3 140 2
LK-STX 14 140 10

dataset. The results of the conditional FP-Tree are

The FP-Growth method can be used in determining
shown in Table 11. The result of the Conditional FP-
the choice of product item pairs. The algorithm focuses
Tree is determined by the calculation of items that meet
on providing a choice of product item pairs based on
the minimum support requirements in the conditional
the results of sales transactions, generating association
pattern base.
rules.
Table 11. Conditional FP-Tree 5) FP-Growth association rules
Item Conditional Pattern Base Conditional
FP-Tree
From the results of itemsets that often appear, the
BKN {BF:3} BF:3 association rules will be formed using (4) so as to
RB {{PS, KN:1}, {CK:3}} CK:3 produce trust values that can be used in decision
BK {RI:1}, {BF:3}, {:5}} BF:3 making by store owners, the results of association rules
CAK {{:2}, {LK, CK, STX:1}, LK:7
{LK, STX:1}, {LK:7}} with 2-itemsets are shown in Table 15.
RI {{LK, STX:5}, {LK :7}, LK: 8, STX: 3
{CK:1}, {:3} Based on Table 15 FP-Growth algorithm obtains
STX {{LK: 14}, {LK, CK :1}, {:1} LK:14

4) Frequent itemset stages Table 14. Support 3-itemset

Item Transaction Total Support
The next step is to form a Frequent Itemset by Transactions (%)
combining sets and subsets of the Conditional FP-Tree LK-STX-CAK 3 140 2
with items. The results of the frequent itemset stage can LK-STX-RI 5 140 4
be seen in Table 12, which is a collection of frequently
Table 15. Association Rules FP-Growth 2-itemset
occurring items that meet the minimum support value No. Rules Support Confidence
of Conditional FP-Tree. The result of frequent items (%) (%)
1 If buying BF → BKN 0.02 0.3
will be found as the minimum value of support using
2 If buying BKN → BF 0.02 1
the set of 2 items and the set of 3 items. To find the 3 If buying CK → RB 0.02 0.18
support value for a set of 2 items using (2), and for 4 If buying RB → CK 0.02 0.75
a set of 3 items using (3), the results of finding the 5 If buying BF → BK 0.02 0.3
6 If buying BK → BF 0.02 0.33
support value for the 2-Itemset are shown in Table 13. 7 If buying LK → CAK 0.06 0.32
8 If buying CAK → LK 0.06 0.8
After a combination of 2-itemset. Furthermore, the 9 If buying STX → CAK 0.02 0.19
combined analysis results of the 2-itemset will proceed 10 If buying CAK → STX 0.02 0.30
to find the relationship with the 3-itemset set before 11 If buying → RI 0.05 0.28
12 If buying RI → LK 0.05 0.58
calculating the association rules to be used in deci- 13 If buying STX → RI 0.02 0.19
sion making by retail owners or related parties, then 14 If buying RI → STX 0.02 0.25
the results of the 3-itemset calculation are shown in 15 If buying LK → STX 0.1 0.56
Table 14. 16 If buying → LK 0.1 0.88

Jurnal Infotel, Vol. 15, No. 4, November 2023

375
https://doi.org/10.20895/infotel.v15i4.980
 ISSN: 2085-3688; e-ISSN: 2460-0997
Implementation of association rule using apriori algorithm and frequent pattern growth for inventory control

16 association rules for 2-itemset, then the calculation LK” and ”If Buying RI then buying LK”. However,
of 3-itemset association rules is carried out using (3) differences also exist between the two. In the apriori
and (4) results are shown in Table 16. algorithm, there is only one rule with the highest
lift ratio, while in the FP-Growth algorithm there are
Table 16. Association Rules FP-Growth 3-itemset
No. Rules Support Confidence two rules with the highest lift ratio. However, in the
(%) (%) apriori algorithm, the lowest lift ratio is 3, while in
1 If buying LK → STXˆCAK 2 12
2 If buying LK → CAKˆSTX 2 12
the FP-Growth algorithm, there are two rules that get
3 If buying LKˆCAK → STX 2 12 the lowest lift with a value of 2. Furthermore, the
4 If buying LKˆSTX → CAK 2 12 calculation of lift values for the 3-itemset hasip is
5 If buying STX → LKˆCAK 2 19 shown in Table 18.
6 If buying STX → CAKˆLK 2 19
7 If buying STXˆCAK → LK 2 19
Both algorithms get the same lift ratio value, this
8 If buying STXˆLK → CAK 2 19
9 If buying CAK → LKˆSTX 2 30 shows that both are equally effective in extracting as-
10 If buying CAK → STXˆLK 2 30 sociation rules from the same dataset. In this case, the
11 If buying CAKˆLK → STX 2 30 analysis does show that the results of both algorithms
12 If buying CAKˆSTX → LK 2 30
13 If buying LK → STXˆRI 4 12
are fully consistent, and there is no difference in the
14 If buying LK → RIˆSTX 4 20 rule-associations generated by apriori and FP-Growth.
15 If buying LKˆSTX → RI 4 20
16 If buying LKˆRI → STX 4 20
17 If buying STX → LKˆRI 4 20 V. C ONCLUSION
18 If buying STX → RIˆLK 4 31
19 If buying STXˆLK → RI 4 31 Based on the research that has been carried out, it
20 If buying STXˆRI → LK 4 31 can be concluded by applying a priori algorithm and
21 If buying RI → LKˆSTX 4 31
FP-Growth has succeeded in finding a combination of
22 If buying RI → STXˆLK 4 42
23 If buying RIˆSTX → LK 4 42 2-itemset and 3-itemsets. The results of the 2-itemset
24 If buying RIˆLK → STX 4 42 analysis found 16 association rules. Meanwhile, 3-
itemsets get 24 association rules. This can be able to
After calculation, 24 association rules were found help store owners place goods according to the criteria
for the 3-itemset of the dataset used. Furthermore, as an of the item.
evaluation of the two algorithms, a lift ratio calculation
will be carried out to determine the extent to which
R EFERENCES
the rules are relevant to the existing data. The lift ratio
is a useful metric to measure the extent to which an [1] E. Wanti, L. P. Maharrani, R. H. A. Prasetya, N. W. Tri-
association is stronger than what is expected. Thus, an pustikasari, and G. N. Ikhtiagung, “Optimation economic order
quantity method for a support system reorder point stock,” Int.
evaluation with a lift ratio will help determine whether J. Electr. Comput. Eng., vol. 10, no. 5, pp. 4992-5000, 2020.
the association rules found are really meaningful and
can be used in further decision-making. [2] K. Nissa and M. T. Siregar, “Analisis pengendalian persediaan
bahan baku kain kemeja poloshirt menggunakan metode eco-
nomic order quantity (EOQ) di PT bina busana internusa,” Int.
IV. D ISCUSSION J. Soc. Sci. Bus., vol. 1, no. 4, pp. 271, 2017.

The calculations that have been carried out using [3] N. N. Merliani, N. I. Khoerida, N. T. Widiawati, L. A. Triana,
and P. Subarkah, “Penerapan algoritma apriori pada transaksi
apriori and FP-Growth algorithms aim to identify the penjualan untuk rekomendasi menu makanan dan minuman,”
rules of significant associations. Furthermore, for the J. Nas. Teknol. dan Sist. Inf., vol. 8, no. 1, pp. 9–16, 2022.
evaluation of the association rules that have been
[4] Fitriah, I. Riadi, and Herman, “Analisis data mining sistem
found, the calculation of the lift ratio value will be car- inventory menggunakan algoritma apriori,” Decod. J. Pendidik.
ried out using (5). The use of lift ratio as an evaluation Teknol. Inf., vol. 3, no. 1, pp. 118–129, 2023.
metric in association analysis is based on fundamental [5] S. E. S. Careza Rizky and Y. Sudarso, “Analisis perbandingan
scientific considerations, namely, to measure the accu- metode EOQ dan metode POQ dengan metode min-max dalam
racy and significance of relationships between items pengendalian persediaan bahan baku pada PT sidomuncul
pupuk nusantara,” Admisi dan Bisnis, vol. 17 No 1, no. ISSN
in association rules. The results of calculating the lift 1411 – 4321, pp. 11–22, 2016.
ratio for the 2-itemset are shown in Table 17.
[6] S. A. Rachmawati, L. Syafirullah, and M. N. Faiz, “Peran-
Results from apriori and FP-Growth algorithms for cangan sistem pengendalian persediaan barang menggunakan
2-itemsets show differences and similarities in the metode EOQ dan ROP berbasis web,” Semin. Nas. Terap. Ris.
Inov. Ke-6, vol. 6, no. 1, pp. 778–786, 2020.
association rules found. Both algorithms produce the
same number of association rules, which is 16 rules. [7] E. Fatma and D. S. Pulungan, “Analisis pengendalian perse-
diaan menggunakan metode probabilistik dengan kebijakan
There is a similarity in the rule with the highest lift backorder dan lost sales,” J. Tek. Ind., vol. 19, no. 1, pp. 38,
ratio, where both find the rule ”If you buy BF, then 2018.
buy BKN” with a lift ratio of 14. In addition, some
[8] S. Bagui, K. Devulapalli, and J. Coffey, “A heuristic approach
other rules with a lift ratio of 5 are also found in for load balancing the FP-growth algorithm on MapReduce,”
both algorithms, such as ”If Buying STX then buying Array, vol. 7, pp. 100035, 2020.

Jurnal Infotel, Vol. 15, No. 4, November 2023

376
https://doi.org/10.20895/infotel.v15i4.980
 ISSN: 2085-3688; e-ISSN: 2460-0997
Implementation of association rule using apriori algorithm and frequent pattern growth for inventory control

Table 17. Ratio 2-itemset Algorithm Apriori and FP-growth

No. Algorithm
Apriori FP-Growth
Rules Lift Rules Lift
Ratio Ratio
1 If buying LK → STX 5 If buying BF → BKN 14
2 If buying STX → LK 5 If buying BKN → BF 14
3 If buying LK → RI 5 If buying CK→ RB 6
4 If buying RI → LK 5 If buying RB → CK 6
5 If buying LK → CAK 4 If buying BF → BK 5
6 If buying CAK → LK 4 If buying BK → BF 5
7 If buying STX → RI 4 If buying LK → CAK 4
8 If buying RI → STX 4 If buying CAK → LK 4
9 If buying STX → CAK 3 If buying STX → CAK 3
10 If buying CAK → STX 3 If buying CAK → STX 3
11 If buying BF → BK 5 If buying LK → RI 3
12 If buying BK → BF 5 If buying RI → LK 3
13 If buying BF → BKN 14 If buying STX → RI 2
14 If buying BKN → BF 4 If buying RI → STX 2
15 If buying CK → RB 6 If buying LK → STX 5
16 If buying RB → CK 6 If buying STX → LK 5

Table 18. Ratio 2-itemset Algorithm Apriori and FP-growth

Apriori FP-Growth
Rules Lift Rules Lift
Ratio Ratio
1 If buying LK → STXˆCAK 6 If buying LK → STXˆCAK 6
2 If buying LK → CAKˆSTX 6 If buying LK → CAKˆSTX 6
3 If buying LKˆCAK → STX 2 If buying LKˆCAK → STX 2
4 If buying LKˆSTX → CAK 1 If buying LKˆSTX → CAK 1
5 If buying STX → LKˆCAK 2 If buying STX → LKˆCAK 2
6 If buying STX → CAKˆLK 3 If buying STX → CAKˆLK 3
7 If buying STXˆCAK → LK 9 If buying STXˆCAK → LK 9
8 If buying STXˆLK → CAK 2 If buying STXˆLK → CAK 2
9 If buying CAK → LKˆSTX 3 If buying CAK → LKˆSTX 3
10 If buying CAK → STXˆLK 5 If buying CAK → STXˆLK 5
11 If buying CAKˆLK → STX 5 If buying CAKˆLK → STX 5
12 If buying CAKˆSTX → LK 15 If buying CAKˆSTX → LK 15
13 If buying LK → STXˆRI 10 If buying LK → STXˆRI 10
14 If buying LK → RIˆSTX 10 If buying LK → RIˆSTX 10
15 If buying LKˆSTX → RI 2 If buying LKˆSTX → RI 2
16 If buying LKˆRI → STX 4 If buying LKˆRI → STX 4
17 If buying STX → LKˆRI 6 If buying STX → LKˆRI 6
18 If buying STX → RIˆLK 6 If buying STX → RIˆLK 6
19 If buying STXˆLK → RI 3 If buying STXˆLK → RI 3
20 If buying STXˆRI → LK 16 If buying STXˆRI → LK 16
21 If buying RI → LKˆSTX 4 If buying RI → LKˆSTX 4
22 If buying RI → STXˆLK 4 If buying RI → STXˆLK 4
23 If buying RIˆSTX → LK 21 If buying RIˆSTX → LK 21
24 If buying RIˆLK → STX 8 If buying RIˆLK → STX 8

[9] L. Linwei, W. Yiping, H. Yepiao, L. Bo, M. Fasheng, and Systems, Services, and Applications (TSSA), 03-04 October
D. Ziqiang, “Optimized apriori algorithm for deformation 2019, Bali, Indonesia, pp. 97–101.
response analysis of landslide hazards,” Comput. Geosci., vol.
170, 2023. [14] M. Fauzy, K. R. Saleh W, and I. Asror, “Penerapan metode
association rule menggunakan algoritma apriori pada simulasi
[10] L. Zheng, “Research on e-commerce potential client mining prediksi hujan wilayah kota Bandung,” J. Ilm. Teknol. Inf.
applied to apriori association rule algorithm,” in Proc. - 2020 Terap., vol. II, no. 2, pp. 221–227, 2016.
Int. Conf. Intell. Transp. Big Data Smart City, ICITBS 2020,
pp. 667–670, 2020. [15] G. G. Prabawa, I. G. M. Darmawiguna, and I. M. A. Wirawan,
“Pengembangan sistem pendukung keputusan pengendalian
[11] W. Thurachon and W. Kreesuradej, “Incremental association persediaan barang menggunakan metode economic order quan-
rule mining with a fast incremental updating frequent pattern tity (EOQ) dan min-max berbasis web (Studi kasus: Apotek
growth algorithm,” IEEE Access, vol. 9, pp. 55726–55741, sahabat qita),” J. Nas. Pendidik. Tek. Inform., vol. 7, no. 2, pp.
2021. 107, 2019.

[12] M. Yasir, M. A. Habib, M. Ashraf, S. Sarwar, M. U. Chaudhry, [16] S. Fauziah and Ratnawati, “Penerapan metode FIFO pada
M. Shahwani, M. Ahmad, CH. M. N. Faisal, “D-GENE: sistem informasi persediaan barang,” J. Tek. Komput., vol. 4,
Deferring the GENEration of power sets for discovering fre- no. 1, pp. 98–108, 2018.
quent itemsets in sparse big data,” IEEE Access, vol. 8, pp.
27375–27392, 2020. [17] E. Callens, “Financial instruments entail liabilities: Ether, bit-
coin, and litecoin do not,” Comput. Law Secur. Rev., vol. 40,
[13] J. R. D. Arcos and A. A. Hernandez, “Efficient apriori algo- Apr 2021.
rithm using enhanced transaction reduction approach,” in 2019
IEEE 13th International Conference on Telecommunication [18] P. Giudici and I. Abu-Hashish, “What determines bitcoin

Jurnal Infotel, Vol. 15, No. 4, November 2023

377
https://doi.org/10.20895/infotel.v15i4.980
 ISSN: 2085-3688; e-ISSN: 2460-0997
Implementation of association rule using apriori algorithm and frequent pattern growth for inventory control

exchange prices? A network VAR approach,” Financ. Res. [38] W. N. Setyo and S. Wardhana, “Implementasi data mining pada
Lett., vol. 28, pp. 309–318, Mar 2019. penjualan produk di CV cahaya etya menggunakan algoritma
fp-growth,” Petir, vol. 12, no. 1, pp. 54–63, 2019,
[19] D. G. Anghel, “A reality check on trading rule performance
in the cryptocurrency market: Machine learning vs. technical [39] Z. Abidin, A. K. Amartya, and A. Nurdin, “Penerapan algo-
analysis,” Financ. Res. Lett., vol. 39, Mar 2021. ritma apriori pada penjualan suku cadang kendaraan roda dua
(studi kasus: Toko prima motor sidomulyo),” J. Teknoinfo, vol.
[20] B. Let, K. Sobanski, W. Swider, and K. Wlosik, “What 16, no. 2, pp. 225, 2022.
drives the popularity of stablecoins? Measuring the frequency
dynamics of connectedness between volatile and stable cryp- [40] E. Irfiani, “Application of apriori algorithms to determine
tocurrencies,” Technol. Forecast. Soc. Change, vol. 189, 2023. associations in outdoor sports equipment stores,” SinkrOn, vol.
3, no. 2, pp. 218, 2019.
[21] S. Zhang and G. Mani, “Popular cryptoassets (bitcoin,
ethereum, and dogecoin), gold, and their relationships: Volatil- [41] L. Ardiantoro and N. Sunarmi, “Badminton player scouting
ity and correlation modeling,” Data Sci. Manag., vol. 4, pp. analysis using frequent pattern growth (FP-growth) algorithm,”
30–39, Des 2021. J. Phys. Conf. Ser., vol. 1456, no. 1, 2020.

[22] L. J. Liebi, “Is there a value premium in cryptoasset markets?,” [42] W. Apridonal M, Y. Choiriah, and A. Akmal, “Penerapan data
Econ. Model., vol. 109, Apr 2022. mining menggunakan metode association rule dengan algo-
ritma apriori untuk analisa pola penjualan barang,” JURTEKSI
[23] A. S. Hoong Lee, L. S. Yap, H. N. Chua, Y. C. Low, and M. (Jurnal Teknol. dan Sist. Informasi), vol. 5, no. 2, pp. 193–198,
A. Ismail, “A data mining approach to analyse crash injury 2019.
severity level,” J. Eng. Sci. Technol., vol. 16, pp. 1–14, 2021.
[43] S. Hu, Q. Liang, H. Qian, J. Weng, W. Zhou, and P. Lin,
[24] S. Wang, J. Cao, and P. S. Yu, “Deep learning for spatio- “Frequent-pattern growth algorithm based association rule min-
temporal data mining: A survey,” IEEE Transactions on Knowl- ing method of public transport travel stability,” Int. J. Sustain.
edge and Data Engineering, vol. 34, no. 8, pp. 1–21, 2019. Transp., vol. 15, no. 11, pp. 879–892, 2021.
[25] A. De Vries, “Cryptocurrencies on the road to sustainability: [44] E. Elisa and N. Azwanti, “Algoritma fp-growth untuk men-
Ethereum paving the way for Bitcoin,” Patterns, vol. 4, no. 1. ganalisa frekuensi pembelian gas elpiji 3 kg,” INTENSIF J.
Cell Press, 13 Januari 2023. Ilm. Penelit. dan Penerapan Teknol. Sist. Inf., vol. 3, no. 1, pp.
69, 2019.
[26] I. Chalkiadakis, A. Zaremba, G. W. Peters, and M. J. Chantler,
“On-chain analytics for sentiment-driven statistical causality in [45] R. M. Anggraeni, “Perbandingan algoritma apriori dan algo-
cryptocurrencies,” Blockchain Res. Appl., vol. 3, no. 2, 2022. ritma fp-growth untuk perekomendasi pada transaksi peminja-
man buku di perpustakaan Universitas Dian Nuswantoro,” Tek.
[27] S. Syahriani, “Penerapan data mining untuk menentukan pola Inform., pp. 1–6, 2014.
penjualan sepatu menggunakan metode algoritma apriori,” Bina
Insa. Ict J., vol. 9, no. 1, pp. 43, 2022.
[28] F. Rahmawati and N. Merlina, “Metode data mining terhadap
data penjualan sparepart mesin fotocopy menggunakan algo-
ritma apriori,” PIKSEL Penelit. Ilmu Komput. Sist. Embed.
Log., vol. 6, no. 1, pp. 9–20, 2018.
[29] Y. A. Ünvan, “Market basket analysis with association rules,”
Commun. Stat. - Theory Methods, vol. 50, no. 7, pp.
1615–1628, 2021.
[30] P. H. Agapito, G. Milano, M. Guzzi and M. Cannataro, “Mining
association rules from disease ontology,” in Proc. - 2019 IEEE
Int. Conf. Bioinforma. Biomed. BIBM 2019, pp. 2239–2243,
2019.
[31] G. Zhang, C. Liu, and M. Tao, “Data mining technology based
on association rules algorithm,” Int. J. Mechatronics Appl.
Mech., vol. 2019, no. 5, pp. 106–112, 2019.
[32] Z. Chen and B. Ye, “Association rule analysis of petrochemical
fire accidents based on the apriori algorithm,” in 2019 9th Int.
Conf. Fire Sci. Fire Prot. Eng. ICFSFPE 2019, pp. 0–4, 2019.
[33] E. Elisa, “Market basket analyis pada mini market ayu dengan
algoritma apriori,” Jurnal Resti, vol. 1, no. 2, 2018.
[34] A. W. O. Gama, I. K. G. D. Putra, and I. P. A. Bayupati,
“Implementasi algoritma apriori untuk menemukan frequent
itemset dalam keranjang belanja,” Maj. Ilm. Teknol. Elektro,
vol. 15, no. 2, pp. 21–26, 2016.
[35] A. Salam, J. Zeniarja, W. Wicaksono, and L. Kharisma, “Pen-
carian pola asosiasi untuk penataan barang dengan menggu-
nakan perbandingan algoritma apriori dan Fp-Growth (study
kasus distro epo store Pemalang),” Dinamik, vol. 23, no. 2, pp.
57–65, 2019.
[36] M. D. Febrianto and A. Supriyanto, “Implementasi algoritma
apriori untuk menentukan pola pembelian produk,” Jurikom,
vol. 9, no. 6, pp. 2010–2020, 2022.
[37] F. AR. Mado, “Analisis persediaan bahan baku produk usaha
sale pisang industri rumah tangga ‘Sofie’ di kota Palu,” e-J.
Agrotekbis, vol. 4, no. 2, pp. 204–209, 2016.

Jurnal Infotel, Vol. 15, No. 4, November 2023

378
https://doi.org/10.20895/infotel.v15i4.980