Ijsrdv1 i2039
0 likes147 views
This document discusses sequential pattern mining, which aims to discover patterns or rules in sequential data where events are ordered by time. It provides background on sequential pattern mining and its applications. The document also discusses related work on mining sequential patterns and rules from time-series data and across multiple sequences. It describes algorithms for efficiently mining sequential patterns at scale from large databases.
![IJSRD - International Journal for Scientific Research & Development| Vol. 1, Issue 2, 2013 | ISSN (online): 2321-0613
All rights reserved by www.ijsrd.com 213
Abstract—Sequential rule mining has been applied in
several domains such as stock market analysis [3], weather
observation and drought management [4].The most famous
approach for sequential rule mining is that of Mannila et al.
[3] and other researchers afterward that aim at discovering
partially ordered sets of events appearing frequently within a
time window in a sequence of events.
I. INTRODUCTION
Recent developments in computing and automation
technologies have resulted in computerizing business and
scientific applications in various areas. Turing the massive
amounts of accumulated information into knowledge is
attracting researchers in numerous domains as well as
databases, machine learning, statistics, and so on. From the
views of information researchers, the stress is on
discovering meaningful patterns hidden in the massive data
sets. Hence, a central issue for knowledge discovery in
databases, additionally the main focus of this thesis, is to
develop economical and scalable mining algorithms as
integrated tools for management systems.
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.
The discovered information will be utilized in
many ways in corresponding applications. For instance,
distinctive the oft times appeared sets of things in a very
retail info will be used to improve the choice creating
of merchandise placement or commercial. Discovering
patterns of client browsing and buying (from either client
records or net traversals) could assist the modeling of user
behaviors for client retention or customized services. Given
the specified databases, whether relational, transactional,
spatial, temporal, or transmission ones, we have a tendency
to could get helpful info once the information discovery
method if acceptable mining techniques square measure
used.
II. BACKGROUND AND PROBLEM DEFINITION
If a collection of data sequences is given, within which
every sequence may be a list of transactions ordered by the
transaction time, the matter of mining sequential patterns [3]
is to get all sequences with a user such minimum support.
Every transaction contains a collection of things. A
sequential pattern is an ordered list (sequence) of item sets.
The item sets that area unit contained within the sequence
area unit referred to as parts of the sequence. For a given
database D that consists of client transactions every group
action consists of the subsequent fields: customer-ID,
transaction-time, and therefore the things purchased within
the group action. an item-set may be a non-empty set of
things, and a sequence is an order list of item-sets. We are
saying a sequence A is contained in another sequence
B if there exists integer’s i1.
Support=
A sequence is an ordered list of elements (transactions).
Each element contains a collection of events (items). Each
element is attributed to a specific time or location. Length of
a sequence, |s|, is given by the number of elements of the
sequence
ID Sequences
1 {1,2},{3},{6},{7},{5}
2 {1,4},{3},{2},(1,2,5,6}
3 {1},{2},{6},{5},{6,7}
4 {2},{6,7},{1,2},{2,3}
Table 1: A Sequence Database
Considering a minimum support = 50% and minimum
confidence = 50%, we get following sequential rules
ID
SEQUENTIAL
RULE
SUPPORT CONFIDENCE
1 {1,2,3} => {5} .5 1.0
2 {1} => {3,5,6} .5 .66
3 {1,2} => {5,6} .75 .75
4 {2} => {5,6} .75 .75
5 {1} => {5,6} .5 .5
.. .. .. ..
Table 2: SEQUENTIAL RULES
sections: sorting phase, finding the massive item-set
phase, transformation section, sequence section, and
greatest phase.
The goal of sequential patterns is to search out the
sequences that have larger than or equal to an explicit user
An Introduction to Effective Sequential Pattern Mining
Shabana Anwar1 Abhishek Raghuvanshi2](https://image.slidesharecdn.com/ijsrdv1i2039-140731042705-phpapp01/85/Ijsrdv1-i2039-1-320.jpg)
![An Introduction to Effective Sequential Pattern Mining
(IJSRD/Vol. 1/Issue 2/2013/0039)
All rights reserved by www.ijsrd.com 214
pre-specified support. Sometimes the method of finding
sequential patterns consists of the subsequent
III. RELATED WORK
As we know, data are changing all the time; especially data
on the web are highly dynamic. As time passes by, new
datasets are inserted; old datasets are deleted while some
other datasets are refreshed. It is transparent that time stamp
is an important attribute of each dataset, also it’s aristocratic
in the process of data mining and it can give us more
accurate and useful information. For example, association
rule mining does not take the time stamp in account, the rule
may Buy A=⇒Buy B. If we take time stamp in account then
we can get more accurate and useful rules such as: Buy A
implies Buy B within two days, three days four days or a
week and a month, or usually people Buy A everyday in a
week. The second kind of rules, business decision can be
more accurate and useful prediction and consequently make
more sound decisions.
However, one important limitation of the algorithms of Das
et al.,[3] and Harms et al. [4] comes from the fact that they
are designed for mining rules occurring frequently in
sequences. As a consequence, these algorithms are
inadequate for discovering rules common to many
sequences. We illustrate this with an example. Consider a
sequence database where each sequence corresponds to a
customer, and each event represents the items bought during
a particular day. Suppose that one wishes to mine sequential
rules that are common to many customers. The algorithms
of Das et al. [3] and Harms et al. [4] are inappropriate since
a rule that appears many times in the same sequence could
have a high support even if it does not appear in any other
sequences. A second example is the application domain of
this paper. We have built an intelligent tutoring agent that
records a sequence of events for each of its executions. We
wish that the tutoring agent discovers sequential rules
between events, common to several of its executions, so that
the agent can thereafter use the rules for prediction during
its following execution.
In order to reduce the number of iterations, the efficient bi-
directional sequential pattern mining approach namely
Recursive Prefix Suffix Pattern detection, RPSP [7]
algorithm is furnished. The RPSP algorithm finds first all
Frequent Itemsets (FI‟s) according to the given minimum
support and transforms the database such that each
transaction is replaced by all the FI‟s it contains and then
finds the patterns. Further the pattern detected based on ith
projected databases, and builds suffix and prefix databases
based on the Apriori properties. Recursive Prefix Suffix
Pattern will increase the number of frequent patterns by
reducing the minimum support and vice versa. Recursion
gets deleted when the detected FI set of prefix or suffix
assigned database of parent database is ineffective. All
patterns that correlate to a particular ith proposition database
of transformed database, that formed into a set, that is dis-
joint from all the other sets. The resultant set of frequent
patterns is the sum of the all disjoint subsets. The proposed
algorithm tested on hypothetical and sequence data and
obtained results were found all satisfactory. Hence, RPSP
algorithm may be applicable to many real world sequential
data sets.
IV. CONCLUSION
We have performed a systematic study on mining of
sequential patterns in large databases and developed a
pattern-growth approach for efficient and scalable mining of
sequential patterns. Instead of refinement of the a priori-like,
candidate generation-and-test approach, such as GSP [3], we
promote a divide-and-conquer approach, called pattern-
growth approach, which is an extension of FP-growth [7], an
efficient pattern-growth algorithm for mining frequent
patterns without candidate generation
REFERENCES
[1] Tan, kumar “introduction to data mining”.
[2] Arun Pujari “ Introduction to data mining”
[3] Han and Kamber, 2000
[4] Das., G., Lin, K.-I., Mannila, H., Renganathan, G., and
Smyth, P. Rule Discovery from Time Series. In Proc. 4th
Int. Conf. on Knowledge Discovery and Data Mining
(New York, USA, August 27-31, 1998), 16-22.
[5] Harms, S. K., Deogun, J. and Tadesse, T. 2002.
Discovering Sequential Association Rules with
Constraints and Time Lags in Multiple Sequences. In
Proc. 13th Int. Symp. on Methodologies for Intelligent
Systems (Lyon, France, June 27-29, 2002), pp. .373-376.
[6] Mannila, H., Toivonen and H., Verkano, A.I. Discovery
of frequent episodes in event sequences. Data Mining
and Knowledge Discovery, 1, 1 (1997), 259-289
[7] Dr P padmaja, P Naga Jyoti, m Bhargava “Recursive
Prefix Suffix Pattern Detection Approach for Mining
Sequential Patterns” IJCA September 2011](https://image.slidesharecdn.com/ijsrdv1i2039-140731042705-phpapp01/85/Ijsrdv1-i2039-2-320.jpg)
Ijsrdv1 i2039
- 1. IJSRD - International Journal for Scientific Research & Development| Vol. 1, Issue 2, 2013 | ISSN (online): 2321-0613 All rights reserved by www.ijsrd.com 213 Abstract—Sequential rule mining has been applied in several domains such as stock market analysis [3], weather observation and drought management [4].The most famous approach for sequential rule mining is that of Mannila et al. [3] and other researchers afterward that aim at discovering partially ordered sets of events appearing frequently within a time window in a sequence of events. I. INTRODUCTION Recent developments in computing and automation technologies have resulted in computerizing business and scientific applications in various areas. Turing the massive amounts of accumulated information into knowledge is attracting researchers in numerous domains as well as databases, machine learning, statistics, and so on. From the views of information researchers, the stress is on discovering meaningful patterns hidden in the massive data sets. Hence, a central issue for knowledge discovery in databases, additionally the main focus of this thesis, is to develop economical and scalable mining algorithms as integrated tools for management systems. 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. The discovered information will be utilized in many ways in corresponding applications. For instance, distinctive the oft times appeared sets of things in a very retail info will be used to improve the choice creating of merchandise placement or commercial. Discovering patterns of client browsing and buying (from either client records or net traversals) could assist the modeling of user behaviors for client retention or customized services. Given the specified databases, whether relational, transactional, spatial, temporal, or transmission ones, we have a tendency to could get helpful info once the information discovery method if acceptable mining techniques square measure used. II. BACKGROUND AND PROBLEM DEFINITION If a collection of data sequences is given, within which every sequence may be a list of transactions ordered by the transaction time, the matter of mining sequential patterns [3] is to get all sequences with a user such minimum support. Every transaction contains a collection of things. A sequential pattern is an ordered list (sequence) of item sets. The item sets that area unit contained within the sequence area unit referred to as parts of the sequence. For a given database D that consists of client transactions every group action consists of the subsequent fields: customer-ID, transaction-time, and therefore the things purchased within the group action. an item-set may be a non-empty set of things, and a sequence is an order list of item-sets. We are saying a sequence A is contained in another sequence B if there exists integer’s i1. Support= A sequence is an ordered list of elements (transactions). Each element contains a collection of events (items). Each element is attributed to a specific time or location. Length of a sequence, |s|, is given by the number of elements of the sequence ID Sequences 1 {1,2},{3},{6},{7},{5} 2 {1,4},{3},{2},(1,2,5,6} 3 {1},{2},{6},{5},{6,7} 4 {2},{6,7},{1,2},{2,3} Table 1: A Sequence Database Considering a minimum support = 50% and minimum confidence = 50%, we get following sequential rules ID SEQUENTIAL RULE SUPPORT CONFIDENCE 1 {1,2,3} => {5} .5 1.0 2 {1} => {3,5,6} .5 .66 3 {1,2} => {5,6} .75 .75 4 {2} => {5,6} .75 .75 5 {1} => {5,6} .5 .5 .. .. .. .. Table 2: SEQUENTIAL RULES sections: sorting phase, finding the massive item-set phase, transformation section, sequence section, and greatest phase. The goal of sequential patterns is to search out the sequences that have larger than or equal to an explicit user An Introduction to Effective Sequential Pattern Mining Shabana Anwar1 Abhishek Raghuvanshi2
- 2. An Introduction to Effective Sequential Pattern Mining (IJSRD/Vol. 1/Issue 2/2013/0039) All rights reserved by www.ijsrd.com 214 pre-specified support. Sometimes the method of finding sequential patterns consists of the subsequent III. RELATED WORK As we know, data are changing all the time; especially data on the web are highly dynamic. As time passes by, new datasets are inserted; old datasets are deleted while some other datasets are refreshed. It is transparent that time stamp is an important attribute of each dataset, also it’s aristocratic in the process of data mining and it can give us more accurate and useful information. For example, association rule mining does not take the time stamp in account, the rule may Buy A=⇒Buy B. If we take time stamp in account then we can get more accurate and useful rules such as: Buy A implies Buy B within two days, three days four days or a week and a month, or usually people Buy A everyday in a week. The second kind of rules, business decision can be more accurate and useful prediction and consequently make more sound decisions. However, one important limitation of the algorithms of Das et al.,[3] and Harms et al. [4] comes from the fact that they are designed for mining rules occurring frequently in sequences. As a consequence, these algorithms are inadequate for discovering rules common to many sequences. We illustrate this with an example. Consider a sequence database where each sequence corresponds to a customer, and each event represents the items bought during a particular day. Suppose that one wishes to mine sequential rules that are common to many customers. The algorithms of Das et al. [3] and Harms et al. [4] are inappropriate since a rule that appears many times in the same sequence could have a high support even if it does not appear in any other sequences. A second example is the application domain of this paper. We have built an intelligent tutoring agent that records a sequence of events for each of its executions. We wish that the tutoring agent discovers sequential rules between events, common to several of its executions, so that the agent can thereafter use the rules for prediction during its following execution. In order to reduce the number of iterations, the efficient bi- directional sequential pattern mining approach namely Recursive Prefix Suffix Pattern detection, RPSP [7] algorithm is furnished. The RPSP algorithm finds first all Frequent Itemsets (FI‟s) according to the given minimum support and transforms the database such that each transaction is replaced by all the FI‟s it contains and then finds the patterns. Further the pattern detected based on ith projected databases, and builds suffix and prefix databases based on the Apriori properties. Recursive Prefix Suffix Pattern will increase the number of frequent patterns by reducing the minimum support and vice versa. Recursion gets deleted when the detected FI set of prefix or suffix assigned database of parent database is ineffective. All patterns that correlate to a particular ith proposition database of transformed database, that formed into a set, that is dis- joint from all the other sets. The resultant set of frequent patterns is the sum of the all disjoint subsets. The proposed algorithm tested on hypothetical and sequence data and obtained results were found all satisfactory. Hence, RPSP algorithm may be applicable to many real world sequential data sets. IV. CONCLUSION We have performed a systematic study on mining of sequential patterns in large databases and developed a pattern-growth approach for efficient and scalable mining of sequential patterns. Instead of refinement of the a priori-like, candidate generation-and-test approach, such as GSP [3], we promote a divide-and-conquer approach, called pattern- growth approach, which is an extension of FP-growth [7], an efficient pattern-growth algorithm for mining frequent patterns without candidate generation REFERENCES [1] Tan, kumar “introduction to data mining”. [2] Arun Pujari “ Introduction to data mining” [3] Han and Kamber, 2000 [4] Das., G., Lin, K.-I., Mannila, H., Renganathan, G., and Smyth, P. Rule Discovery from Time Series. In Proc. 4th Int. Conf. on Knowledge Discovery and Data Mining (New York, USA, August 27-31, 1998), 16-22. [5] Harms, S. K., Deogun, J. and Tadesse, T. 2002. Discovering Sequential Association Rules with Constraints and Time Lags in Multiple Sequences. In Proc. 13th Int. Symp. on Methodologies for Intelligent Systems (Lyon, France, June 27-29, 2002), pp. .373-376. [6] Mannila, H., Toivonen and H., Verkano, A.I. Discovery of frequent episodes in event sequences. Data Mining and Knowledge Discovery, 1, 1 (1997), 259-289 [7] Dr P padmaja, P Naga Jyoti, m Bhargava “Recursive Prefix Suffix Pattern Detection Approach for Mining Sequential Patterns” IJCA September 2011