1

Apriori Algorithm

A Review Paper of Association Rule Mining

Using Apriori Algorithm

Abstract— Association rule mining is a II. METHODOLOGY

fundamental data mining technique that attempts to
discover formerly unknown patterns and The review takes into consideration ten carefully
relationships among data sets. The Apriori selected research articles representing work in
algorithm is the most popular and widely applied relation to application, optimizing, and utilizing the
algorithm to derive association rules. In this review algorithm of Apriori. All such articles have been
paper, ten research articles are examined to analyzed for contribution towards new approaches
determine the progress, applications, challenges, development, algorithm application in numerous
and future prospect of association rule mining based areas, and overcoming current restrictions.
on the Apriori algorithm. The paper provides an
overview of methods, advancements, and real III. CONCEPTS
implementations and identifies gaps in the current
literature. 1) Association Rule Mining
Association rule mining is a form of mining for
finding useful relations, patterns, and associations
Index Terms— Association rule mining, Apriori between items in a big dataset. Rules occur most
algorithm, Data mining, Frequent itemsets, Market often in form, and and and represent sets of items
basket analysis, Algorithm optimization, and transactions and sets, respectively. Metrics such
Computational efficiency, Big data analytics, as support, confidence, and lift evaluate such forms
Parallel computing, Real-world applications of rules, such as in proportion of transactions in a
big dataset with and, confidence defining
I. INTRODUCTION conditional probability for appearing in a
transaction when in a transaction, and lift, a metric
Association rule mining is one of the most
for times as probable when in a transaction versus
important processes involved in data mining,
presence when not in a transaction with.
intended with an objective to extract significant
Association rule mining aims to produce rules with
relations between variables in enormous datasets. predeterminate values for confidence and support
Formed by Rakesh Agrawal and Ramakrishnan such that its usability and effectiveness can be
Srikant [1], one of the oldest algorithms in its assured.
family, Apriori utilizes breadth-first search for
2 The Apriori Algorithm
searching for frequent sets and creating association
rules. Despite its general use, its efficiency and
The Apriori algorithm is a simple algorithm for
performance become cumbersome when dealing
mining association rules and stems from the
with enormous datasets. In its critique, in this property of Apriori, according to which a frequent
article, an overview of current work, improvements, itemset will have all its subsets being frequent items
implementations, and weaknesses of the algorithm as well. It works iteratively in discovering frequent
Apriori will be discussed itemsets in a form first creating a candidate of a
specific size, its instances in a dataset being
counted, rejecting candidates failing minimum-
 2
Apriori Algorithm

support, and then recursively attempting for

itemsets till no candidates. Frequent itemsets
produced then become utilized in producing IV. LITERATURE REVIEW
association rules.The Apriori algorithm is simple 1) Improvements to the Apriori Algorithm
and transparent but computationally costly with
many reads in a database and high volumes of The Apriori algorithm faces computational
produced candidates, particularly in high- challenges which multiple researchers have worked
dimensional databasesdimensionality to enhance through their proposed improvements.
3 Frequent Itemsets An optimized Apriori framework developed by
Mohammed Al-Maolegi and Bassam Arkok [2]
Frequent itemsets, which form a basis for mining
decreases the number of scans needed over the
association rules, constitute sets of items whose
dataset thus enhancing performance. Researchers
occurrences in transactions occur over a minimum
predetermined value of support. Frequent itemsets introduced two memory-saving strategies that
have significant value and form a basis for combined either hash-based structures or tree-based
generating association rules, but mining them in an approaches.
efficient form poses a significant challenge,
particularly in high-volume datasets. Several 2 Applications in Real-World Scenarios
proposed improvements of the Apriori algorithm
have been proposed in an effort to address such a
challenge. For instance, tree structures and hash The Apriori algorithm has been applied in various
structures have been proposed in an effort to utilize fields, including:
efficient use of memory and reduce computational
cost. Methods for transactional reduction, with a • Retail and E-commerce: Transaction data
target of minimizing transactions under an iterative analysis identifies common item groupings which
scan, have also been proposed [7]. Parallel and helps in market basket analysis [3].
distribution computational methodologies have, in
addition, been adopted with an aim of enhancing • Healthcare: The identification of symptom-
algorithm scalability [8]. diagnosis patterns helps healthcare practitioners
advance patient treatment methods [4].
4 Variants and Extensions
• Education: Student performance records function
To increase its efficiency and make it suitable for as mining data to build customized educational
certain applications, variations and modifications of methods [5].
the Apriori algorithm have been developed. More
efficient versions of the algorithm combines the The Apriori algorithm applies to sales data analyses
hash based pruning and tree based pruning both to through H. Y. Xie's [6] study which demonstrated
reduce the search space of candidate itemsets and to inventory optimization..
increase the speed of the process. Modern
computing devices, because of the widespread use
of multicore processors, have brought about the 3 Comparative Analyses
need for parallel and distributed versions of the
algorithm. These shifts show the applicability of the
Research examined the Apriori algorithm through
Apriori algorithm for modern data mining
comparisons with alternative approaches that
requirements [10].
included Frequent Pattern (FP)-Growth algorithm.
Research showed Apriori's simplicity together with
its interpretability while highlighting its poor
 3
Apriori Algorithm

performance when dealing with large dense Engineering Research, vol. 4, no. 5, pp. 1-5, 2013.
datasets. [7]. Available:
https://www.ijser.org/researchpaper/Advanced-
Apriori-Algorithms.pdf

V. DISCUSSION AND ANALYSIS [4] D. S. Patil and S. S. Joshi, "Educational Data

Mining using Improved Apriori Algorithm,"
International Journal of Computer Science and
Research in this field focuses on three main
Technology, vol. 3, no. 5, pp. 1-4, 2014. Available:
optimization approaches which include parallel and
https://www.ripublication.com/irph/ijict_spl/08_ijic
distributed computing [8] and hybrid machine
tv3n5spl.pdf
learning with Apriori techniques [9] alongside
customizations for specific domains [10]. The [5] V. J. Dhole and N. P. Patil, "Study of Various
Apriori algorithm shows strengths yet deals with Improved Apriori Algorithms," IOSR Journal of
two main weaknesses including high operational Computer Engineering, vol. 15, no. 3, pp. 55-58,
expenses and poor capability to process sparse data 2013. Available: https://www.iosrjournals.org/iosr-
collections [7]. Research today fails to present a jce/papers/conf.15013/Volume%202/13.%2055-
single unifying framework for adapting Apriori 58.pdf
across various datasets while showing minimal
exploration of its potential applications in IoT and [6] R. Agrawal and R. Srikant, "A New Improved
big data analytics [4]. Apriori Algorithm for Association Rules Mining,"
International Journal of Engineering Research and
VI. CONCLUSION Technology, vol. 2, no. 6, pp. 1-6, 2013. Available:
The analysis presents important aspects of the https://www.ijert.org/research/a-new-improved-
apriori-algorithm-for-association-rules-mining-
Apriori algorithm which serves multiple uses in
IJERTV2IS60844.pdf
association rule mining applications. Data mining
relies upon the Apriori algorithm as a fundamental [7] S. R. Joshi and K. M. Desai, "Improving
practice yet needs continuous research to overcome Efficiency of Apriori Algorithm Using Transaction
known drawbacks and expand real-world usage. Reduction," International Journal of Scientific and
Expansive research directions integrate Research Publications, vol. 3, no. 1, pp. 1-5, 2013.
sophisticated computational methods together with Available: https://www.ijsrp.org/research-paper-
fresh explorations of unique applications operating 1301/ijsrp-p1397.pdf
within shifting data environments of growing
[8] P. K. Reddy, "Performance Study of Distributed
dimensionality.
Apriori-like Frequent Itemsets Mining," arXiv
preprint, 2019. Available: [https://ar
References
[9] T. Ahmed and S. Goyal, "A Prefixed-Itemset-
[1] R. Agrawal and R. Srikant, "Fast Algorithms for
Based Improvement for Apriori Algorithm," arXiv
Mining Association Rules," Proceedings of the 20th
preprint, 2016. Available:
International Conference on Very Large Data
https://arxiv.org/abs/1601.01746
Bases, 1994.
[10] S. R. Joshi and K. N. Mahajan, "Optimizing
[2] M. Al-Maolegi and B. Arkok, "An Improved
the Apriori Algorithm for Frequent Itemset
Apriori Algorithm for Association Rules," arXiv
Mining," International Journal of Engineering
preprint, 2014. Available:
Research and Applications (IJERA), vol. 2, no. 5,
https://arxiv.org/pdf/1403.3948
pp. 456-460, 2012. Available:
[3] N. Soni and S. Srivastava, "Advanced Apriori http://www.ijera.com/papers/Vol2_issue5/C025456
Algorithms," International Journal of Scientific and 460.pdf