Fast Sequential Rule Mining
0 likes192 views
This paper proposes a novel method for fast sequential rule mining, which is more efficient than existing approaches by utilizing a pattern-growth method rather than a generate-candidate-and-test approach. The proposed algorithm identifies valuable sequential patterns through a systematic process involving itemset generation and support confidence calculation, ultimately demonstrating superior performance compared to previous algorithms. It highlights the application of sequential pattern mining in various fields, including business and biology, while addressing challenges related to large data sets.
![IJSRD - International Journal for Scientific Research & Development| Vol. 2, Issue 08, 2014 | ISSN (online): 2321-0613
All rights reserved by www.ijsrd.com 170
Fast Sequential Rule Mining
Bharat Parmar1
Prof. Gajendra Singh Chandel2
1,2
Department of Computer Science
1,2
Sri Satya Sai Institute of Science and Technology, Indore, Bhopal, India
Abstract— In this paper, we have proposed a novel
sequential mining method. The method is fast in
comparison to existing method. Data mining, that is
additionally cited as knowledge discovery in databases,
has been recognized because the method of extracting
non-trivial, implicit, antecedently unknown, and probably
helpful data from knowledge in databases. The
information employed in the mining method usually
contains massive amounts of knowledge collected by
computerized applications. As an example, bar-code
readers in retail stores, digital sensors in scientific
experiments, and alternative automation tools in
engineering typically generate tremendous knowledge into
databases in no time. Not to mention the natively
computing- centric environments like internet access logs
in net applications. These databases therefore work as
rich and reliable sources for information generation
and verification. Meanwhile, the massive databases give
challenges for effective approaches for information
discovery.
Key words: simple current mirror, Low voltage current
mirror, level shifted Current mirror, Level shifted low
voltage current mirror, Dynamic range
I. SEQUENTIAL RULE MINING
The sequential pattern is a sequence of itemsets that
regularly occurred in an exceedingly specific order, all
items within the same itemset are supposed to have
constant transaction duration or inside a time-
gap. Sequential patterns indicate the correlation between
transactions whereas association rules represent intra-
transaction relationships.
Sequential pattern mining was initially introduced
by Agrawal and Srikant [1]. It's the process of extracting
certain sequential patterns whose support exceeds a
predefined lowest support threshold. Since the number of
sequences may be very massive, and users have completely
different interests and necessities, to induce the most
interesting sequential patterns sometimes a minimum
support is predefined by the users. By using the minimum
support we will prune out those sequential patterns of no
interest, consequently creating the mining method more
economical. Clearly a better support of a sequential
pattern is desired for additional helpful and interesting
sequential patterns. But some sequential patterns that don't
satisfy the support threshold are still fascinating. Yang,
Wang & Yu (2001) [3] introduced another metric
known as surprise to measure the powerfulness of
sequences. A sequence s is a stunning pattern if its
occurrence differs greatly from the expected occurrence,
once all items are treated equally. Within the surprise
metric the information gain was proposed to measure the
degree of surprise, as detailed by yang [3].
Sequential pattern mining is used in a very great
spectrum of areas. In machine biology, sequential pattern
mining is employed to analyses the mutation patterns of
various amino acids. Business organizations use
sequential pattern mining to check client behaviors.
Sequential pattern mining is additionally utilized in
system performance analysis and telecommunication
network analysis.
General algorithms for sequential pattern mining
are proposed recently. Most of the essential and earlier
algorithms for sequential pattern mining are based on
the Apriori property proposed by Srikant and Agrawal
[2]. The property states that any sub-pattern of a frequent
pattern should be frequent. Based on this heuristic, a
series of Apriori-like algorithms are proposed: AprioriAll,
Apriori Some, Dynamic some [2], GSP [4] and SPADE
[5]. Later, a series of information projection primarily
based algorithms were proposed , including
FreeSpan [6] and PrefixSpan [7]. By using the minimum
support we will prune out those sequential patterns of no
interest, consequently creating the mining method more
economical. Many closed sequence mining algorithmic
rule were additionally i n t r o d u c e d , l i k e C l o S p a n
[8] a n d T S P [ 9]. Most o f t h o s e projection-based
algorithms are explained within the next subsections.
II. PROPOSED SOLUTION
A source database D.
MST (Minimum Support Threshold).
MCT (Minimum Confidence Threshold).
III. PROCEDURE
We set the minimum support and scan the database
to get 1-itemsets. In this step, we also count each
item’s support by using compressed data structure,
i.e. head and body of the database. Here body of
the database contain itemset with their support and
arranges in the lexicographic order, i.e. sorted
order. The proposed method first scans the
sequential data base once & it count the support
single size frequent items. the algorithm then
arrange each pair of frequent elements in form of a
rule. Like for a pair {1,2}, it generates two rules as
: 1 => 2 & 2=>1. It also eliminates all infrequent
items.
Then, the sequential support and sequential
confidence of every rule is calculated. All rules
having support and confidence greater than the
minimum threshold are valid rules.
Then all the rules found in step 2 are expanded on
their left and right side as follows:
Left-side growth is the method of
taking two rules X⇒Y and Z⇒Y,
wherever X and Z are itemsets of size n
sharing n-1 items, generate a larger rule
X∪Z ⇒Y. Right-side expansion is the
process of taking two rules Y⇒X and
Y⇒Z, wherever X and Z are unit itemets](https://image.slidesharecdn.com/ijsrdv2i8033-150923091606-lva1-app6891/85/Fast-Sequential-Rule-Mining-1-320.jpg)
![Fast Sequential Rule Mining
(IJSRD/Vol. 2/Issue 08/2014/041)
All rights reserved by www.ijsrd.com 171
of size n sharing n-1 items, to generate a
larger rule Y⇒X∪Z. These two
processes are applied recursively to seek
out all rules ranging from rule of size 1*1
(for example, rules of size 1*1 permits
finding rules of size 1*2 and rule of size
2*1).
IV. CONCLUSION
In this paper, we tend to conferred a completely unique
algorithm for mining sequential rules common to several
sequences. Unlike previous algorithms, it doesn't use a
generate-candidate-and-test approach. Instead, it uses a
pattern-growth approach for discovering valid rules
specified it will be rather more efficient and scalable.
It initial finds rules between 2 items and so
recursively grows them by scanning the database for
single items that would expand their left or right
components. we've evaluated the performance of our
algorithm by scrutiny it with CMRules algorithms.
Results show that our algorithm clearly outperforms
CMRules and incorporates a higher scalability.
REFERENCES
[1] Rakesh Agr a wa l , & Ramakrishnan
Srika nt, 1995, Mini ng Sequential
Patterns. Proc. Int. Conf. on Data Engineering, pp.
3-14.
[2] Rakesh Agrawal, and Ramakrishnan Srikant,
1994. “Fast Algorithms for Mining Association
Rules”, In Proceedings of the 20th Int. Conf.
Very Large Data Bases, pp. 487-499.
[3] Yang, J., Wang, W., & Yu, P. S. (2001).
Infominer: mining surprising periodic patterns.
Proceedings of the Seventh ACM SIGKDD
International Conference on Knowledge Discovery
and Data Mining.
[4] Srikant, R., & Agrawal, R. (1996).Mining
Sequential Patterns: Generalizations and
Performance Improvements. Proceedings of the
5th International Conference on Extending
Database Technology: Advances in Database
Technology.
[5] Zaki, M. J. (2001). SPADE: An Efficient
Algorithm for Mining Frequent Sequences. Journal
of Machine Learning, 42(1-2), 31-60.
[6] Han, J., Pei, J., Mortazavi-Asl, B., Chen, Q.,
Dayal, U., & Hsu, M.-C. (2000). FreeSpan:
frequent pattern-projected sequential pattern
mining. Proceedings of the Sixth ACM
SIGKDD International Conference on
Knowledge Discovery and Data Mining.
[7] Pei, J., Han, J., Mortazavi-Asl, B., Pinto, H.,
Chen, Q., Dayal, U., et al. (2001). PrefixSpan:
Mining Sequential Patterns by Prefix-
Projected Growth. Proceedings of the 17th
International Conference on Data Engineering.
[8] Yan, X., Han, J., & Afshar, R. (2003). CloSpan:
Mining Closed Sequential Patterns in Large
Datasets. Proceedings of the 2003 SIAM
International Conference on Data Mining
(SDM'03).
[9] Tzvetkov, P., Yan, X., & Han, J. (2005). TSP:
Mining top-k closed sequential patterns.
Knowledge and Information Systems, 7(4), 438-457.](https://image.slidesharecdn.com/ijsrdv2i8033-150923091606-lva1-app6891/85/Fast-Sequential-Rule-Mining-2-320.jpg)
Fast Sequential Rule Mining
- 1. IJSRD - International Journal for Scientific Research & Development| Vol. 2, Issue 08, 2014 | ISSN (online): 2321-0613 All rights reserved by www.ijsrd.com 170 Fast Sequential Rule Mining Bharat Parmar1 Prof. Gajendra Singh Chandel2 1,2 Department of Computer Science 1,2 Sri Satya Sai Institute of Science and Technology, Indore, Bhopal, India Abstract— In this paper, we have proposed a novel sequential mining method. The method is fast in comparison to existing method. Data mining, that is additionally cited as knowledge discovery in databases, has been recognized because the method of extracting non-trivial, implicit, antecedently unknown, and probably helpful data from knowledge in databases. The information employed in the mining method usually contains massive amounts of knowledge collected by computerized applications. As an example, bar-code readers in retail stores, digital sensors in scientific experiments, and alternative automation tools in engineering typically generate tremendous knowledge into databases in no time. Not to mention the natively computing- centric environments like internet access logs in net applications. These databases therefore work as rich and reliable sources for information generation and verification. Meanwhile, the massive databases give challenges for effective approaches for information discovery. Key words: simple current mirror, Low voltage current mirror, level shifted Current mirror, Level shifted low voltage current mirror, Dynamic range I. SEQUENTIAL RULE MINING The sequential pattern is a sequence of itemsets that regularly occurred in an exceedingly specific order, all items within the same itemset are supposed to have constant transaction duration or inside a time- gap. Sequential patterns indicate the correlation between transactions whereas association rules represent intra- transaction relationships. Sequential pattern mining was initially introduced by Agrawal and Srikant [1]. It's the process of extracting certain sequential patterns whose support exceeds a predefined lowest support threshold. Since the number of sequences may be very massive, and users have completely different interests and necessities, to induce the most interesting sequential patterns sometimes a minimum support is predefined by the users. By using the minimum support we will prune out those sequential patterns of no interest, consequently creating the mining method more economical. Clearly a better support of a sequential pattern is desired for additional helpful and interesting sequential patterns. But some sequential patterns that don't satisfy the support threshold are still fascinating. Yang, Wang & Yu (2001) [3] introduced another metric known as surprise to measure the powerfulness of sequences. A sequence s is a stunning pattern if its occurrence differs greatly from the expected occurrence, once all items are treated equally. Within the surprise metric the information gain was proposed to measure the degree of surprise, as detailed by yang [3]. Sequential pattern mining is used in a very great spectrum of areas. In machine biology, sequential pattern mining is employed to analyses the mutation patterns of various amino acids. Business organizations use sequential pattern mining to check client behaviors. Sequential pattern mining is additionally utilized in system performance analysis and telecommunication network analysis. General algorithms for sequential pattern mining are proposed recently. Most of the essential and earlier algorithms for sequential pattern mining are based on the Apriori property proposed by Srikant and Agrawal [2]. The property states that any sub-pattern of a frequent pattern should be frequent. Based on this heuristic, a series of Apriori-like algorithms are proposed: AprioriAll, Apriori Some, Dynamic some [2], GSP [4] and SPADE [5]. Later, a series of information projection primarily based algorithms were proposed , including FreeSpan [6] and PrefixSpan [7]. By using the minimum support we will prune out those sequential patterns of no interest, consequently creating the mining method more economical. Many closed sequence mining algorithmic rule were additionally i n t r o d u c e d , l i k e C l o S p a n [8] a n d T S P [ 9]. Most o f t h o s e projection-based algorithms are explained within the next subsections. II. PROPOSED SOLUTION A source database D. MST (Minimum Support Threshold). MCT (Minimum Confidence Threshold). III. PROCEDURE We set the minimum support and scan the database to get 1-itemsets. In this step, we also count each item’s support by using compressed data structure, i.e. head and body of the database. Here body of the database contain itemset with their support and arranges in the lexicographic order, i.e. sorted order. The proposed method first scans the sequential data base once & it count the support single size frequent items. the algorithm then arrange each pair of frequent elements in form of a rule. Like for a pair {1,2}, it generates two rules as : 1 => 2 & 2=>1. It also eliminates all infrequent items. Then, the sequential support and sequential confidence of every rule is calculated. All rules having support and confidence greater than the minimum threshold are valid rules. Then all the rules found in step 2 are expanded on their left and right side as follows: Left-side growth is the method of taking two rules X⇒Y and Z⇒Y, wherever X and Z are itemsets of size n sharing n-1 items, generate a larger rule X∪Z ⇒Y. Right-side expansion is the process of taking two rules Y⇒X and Y⇒Z, wherever X and Z are unit itemets
- 2. Fast Sequential Rule Mining (IJSRD/Vol. 2/Issue 08/2014/041) All rights reserved by www.ijsrd.com 171 of size n sharing n-1 items, to generate a larger rule Y⇒X∪Z. These two processes are applied recursively to seek out all rules ranging from rule of size 1*1 (for example, rules of size 1*1 permits finding rules of size 1*2 and rule of size 2*1). IV. CONCLUSION In this paper, we tend to conferred a completely unique algorithm for mining sequential rules common to several sequences. Unlike previous algorithms, it doesn't use a generate-candidate-and-test approach. Instead, it uses a pattern-growth approach for discovering valid rules specified it will be rather more efficient and scalable. It initial finds rules between 2 items and so recursively grows them by scanning the database for single items that would expand their left or right components. we've evaluated the performance of our algorithm by scrutiny it with CMRules algorithms. Results show that our algorithm clearly outperforms CMRules and incorporates a higher scalability. REFERENCES [1] Rakesh Agr a wa l , & Ramakrishnan Srika nt, 1995, Mini ng Sequential Patterns. Proc. Int. Conf. on Data Engineering, pp. 3-14. [2] Rakesh Agrawal, and Ramakrishnan Srikant, 1994. “Fast Algorithms for Mining Association Rules”, In Proceedings of the 20th Int. Conf. Very Large Data Bases, pp. 487-499. [3] Yang, J., Wang, W., & Yu, P. S. (2001). Infominer: mining surprising periodic patterns. Proceedings of the Seventh ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. [4] Srikant, R., & Agrawal, R. (1996).Mining Sequential Patterns: Generalizations and Performance Improvements. Proceedings of the 5th International Conference on Extending Database Technology: Advances in Database Technology. [5] Zaki, M. J. (2001). SPADE: An Efficient Algorithm for Mining Frequent Sequences. Journal of Machine Learning, 42(1-2), 31-60. [6] Han, J., Pei, J., Mortazavi-Asl, B., Chen, Q., Dayal, U., & Hsu, M.-C. (2000). FreeSpan: frequent pattern-projected sequential pattern mining. Proceedings of the Sixth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. [7] Pei, J., Han, J., Mortazavi-Asl, B., Pinto, H., Chen, Q., Dayal, U., et al. (2001). PrefixSpan: Mining Sequential Patterns by Prefix- Projected Growth. Proceedings of the 17th International Conference on Data Engineering. [8] Yan, X., Han, J., & Afshar, R. (2003). CloSpan: Mining Closed Sequential Patterns in Large Datasets. Proceedings of the 2003 SIAM International Conference on Data Mining (SDM'03). [9] Tzvetkov, P., Yan, X., & Han, J. (2005). TSP: Mining top-k closed sequential patterns. Knowledge and Information Systems, 7(4), 438-457.