Download Finding Frequent and Maximal Periodic Patterns in

International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 1, January 2014)
Finding Frequent and Maximal Periodic Patterns in
Spatiotemporal Databases towards Big Data
O.Obulesu1, A.Rama Mohan Reddy2, M.Thrilok Reddy3
1,3
Asst. Professor, Department of IT, SVEC, JNTUA, A.P.
2
Professor, Dept of CSE, SVUCE, Tirupati
The periodicity detection is a process of finding
temporal regularities within the Time-Series and the goal
of analyzing a Time-Series Database is to find how
frequent a periodic pattern (full or partial) is repeated
within time intervals. In general, there are three types of
periodic patterns can be detected in a time series Database
such as Symbol Periodicity, Sequence Periodicity or
Partial Periodic Patterns and Segment or Full-Cycle
Periodicity [22]. We consider a set of Boolean SpatioTemporal (ST) event types such as crime types and their
instances. The Boolean event types are important because
primary concern is the occurrence or absence of an event
type at a particular location and time. Given such a set,
cascading spatio-temporal patterns are partially ordered
subsets of event types whose instances are spatial
neighbours and occur in a series of stages. It is an
example of a Cascading Spatiotemporal Pattern Discovery
CSTP [24]. The first event1 in the CSTP is the hurricane
event. Subsequent events are represented by events such
as heavy rainfall, localized flooding and wind damage
creates a CSTP from an urban crime dataset. The first
event here is bar-closings (represented by circles), and
subsequent events are assaults (represented by triangles)
and drunk driving (represented by squares). It shows
individual instances of events of bar-closing, assault and
drunk driving with their location and time. It shows the
locations of all event instances. The problem is 80-90% of
data produced today is unstructured and Gartner Predicts
800% data growth over next 5 years [26]. Big data is a
term applied to data sets whose size is beyond the ability
of commonly used software tools to capture, manage, and
process the data within a tolerable elapsed time.
Abstract— Data mining used to find hidden knowledge
from large amount of Databases. Periodic Pattern Mining is
useful in Weather Forecasting, Fraud Detection and GIS
Applications. In General, spatio-temporal pattern discovery
process finds the partially ordered subsets of the eventtypes whose instances are located together and occur
serially for a given collection of Boolean spatio-temporal
event-types. Big Data concerns large-volume, complex,
growing data sets with multiple, autonomous sources. With
the fast development of networking, data storage, and the
data collection capacity, Big Data is now rapidly expanding
in all science and engineering domains, including physical,
biological and bio-medical sciences. In this paper, a new
framework is proposed to find spatiotemporal patterns
from Big Data. Existing algorithms are well in computation
of necessary patterns, but more problematic when they are
applied to Big Data. Big Data is a new trend used to analyze
the datasets that due to their large size and complexity,
Developers cannot manage them with traditional current
algorithms or data mining software tools. Big Data mining
is the capability of extracting useful information from these
large datasets or streams of data, that due to its volume,
variety, and velocity, it was not possible before to do it. The
Big Data challenge is becoming one of the most exciting
opportunities for the next years. This Paper focuses on a
broad overview of pattern mining algorithms and
significance in Spatiotemporal Databases, its current status,
trade-offs, and forecast to the big data pattern mining
future.
Keywords — Periodicity Detection, Spatial Patterns, Big
Data, Cascading Spatiotemporal Pattern Discovery, MapR
I. INTRODUCTION
Data mining is a useful tool for extracting nontrivial,
implicit, previously unknown and potentially useful
information or pattern from large databases [28].It is the
process of applying the methods like neural networks,
clustering, genetic algorithms (1950s), decision trees
(1960s) and support vector machines (1980s) to data with
the intention of uncovering hidden patterns. A TimeSeries Database is a collection of data values gathered
generally at uniform interval of time to reflect certain
behavior of an entity. In Real World, There are several
examples of Time-Series such as Weather Conditions of a
particular location, Spending Patterns, Stock Growth, and
Transactions in a Supermarket, Network Delays, Power
Consumption, Computer Network Fault Analysis and
Security Breach detection, Earthquake Prediction.
II. BIG DATA PATTERN MINING
This section gives motivation and directions for
pattern mining towards Big Data research. The basic
three parameters of Big Data are volume, variety, and
velocity of information, drives unprecedented complexity
and opportunity [25]. The main idea of this paper is to
introduce periodic pattern mining in big data to be useful
in rare event analysis converted as negative patterns
including positive patterns [34]. Data and content
increase over coming decade. Most of the data generated
from web is unstructured. In 2025, 35 Zettabytes of data
to be generated.
266
International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 1, January 2014)
Today the storage levels of data in the forms of
Gigabytes, Terabytes, Petabytes, Exabytes, Zettabytes
and Yottabytes only.
Again in 1995, same authors proposed two new
algorithms such as AprioriSome and AprioriAll to mine
sequential patterns from a large database of customer
transactions, where each transaction consists of
customer-id, transaction time, and the items bought in the
transaction [2].
In 1998, Roberto J. Bayardo Jr. Proposed a new
pattern mining algorithm called the Max-Miner, that
scales roughly linearly in the number of maximal patterns
embedded in a database irrespective of the length of the
longest pattern [3]. Max-Miner is also easily made to
incorporate additional constraints on the set of frequent
itemsets identified. Later, Garofalakis et al. in 1999,
proposed SPIRIT, a method of mining user-specified
sequential patterns by using regular expression
constraints [35]. In 2000, J. Han, J. Pei and R. Mao
proposed an efficient CLOSET Algorithm, for mining
closed itemsets with the help of three techniques such as
FP-Tree structure, Single prefix path compression and
partition based projection mechanism. It is efficient and
scalable over large databases, and faster than the
traditional algorithms [4].
In 2000, Yves Bastide, Rafik Taouil, Nicolas Pasquier,
Gerd Stumme proposed the algorithm PASCAL which
introduces a novel optimization of the well-known
algorithm Apriori and it is based on pattern counting
inference that relies on the concept of key patterns [5].
PASCAL is better than three algorithms such as Apriori,
Close and Max-Miner and it is the most efficient
algorithm for mining frequent patterns. Later, Jiawei
Han, Jian Pei, and Yiwen Yin proposed a novel frequent
pattern tree (FP-tree) structure, which is an extended
prefix tree structure for storing compressed, crucial
information about frequent patterns, and develop an
efficient FP-tree based mining method, FP-growth, for
mining the complete set of frequent patterns by pattern
fragment growth [6].
Later in 2001, Doug Burdick, Manuel Calimlim,
Johannes Gehrke proposed a new algorithm called
MAFIA, for mining maximal frequent itemsets from a
transactional database [7]. This algorithm outperforms
previous work by a factor of three to five. Again in 2001,
Ilias Tsoukatos and Dimitrios Gunopulos solved the
problem of mining spatiotemporal patterns and finding
sequences of events that occur frequently in
spatiotemporal datasets with the help of DFS_MINE, a
new algorithm for fast mining of frequent spatiotemporal
patterns in environmental data, which has the advantage
that the amount of space required is minimal [8].
2.1 Pattern Mining Classification:
The following different patterns to be useful in further
research in Data Mining domain [28].
(a) Kinds of patterns and rules
(b) Mining Methods
(c) Extensions and applications
(a) Kinds of Patterns and Rules:
i) Basic Patterns – Frequent patterns, Association
rules and Closed/maximal Patterns etc.
ii) Multilevel and Multidimensional Patterns –
Uniform and Item-set based High dimensional
patterns etc.
iii) Extended Patterns – Approximate, Uncertain,
Compressed, Rare/negative and Colossal Patterns
etc.
(b) Mining Methods:
i) Basic Mining methods – Candidate Generation,
Pattern Growth, Vertical format etc.
ii) Mining interesting patterns – Interestingness,
Constraint-based, Correlation Rules and Exception
Rules etc.
iii) Distributed, parallel & Incremental – Stream
Patterns, Distributed/Parallel and Incremental
Mining etc.
(c) Extensions and Applications:
i) Extended Data types – Sequential & Time-Series,
Structural, Spatial, Temporal and Multimedia
patterns.
ii) Applications – patterns based Classification &
Clustering, Semantic Annotation and Privacy
Preserving.
III. RELATED WORK
In 1994, Rakesh Agrawal and Ramakrishnan Srikanth
identified the problem of discovering association rules
between items in a large database of sales transactions
and proposed two new algorithms for solving this
problem that are fundamentally different from the known
algorithms such as AprioirHybrid [1]. AprioriHybrid is a
combination of two algorithms such as Apriori and
AprioriTid also has excellent scale-up properties with
respect to the transaction size and the number of items in
the database.
267
International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 1, January 2014)
In 2002, Mohammed J. Zaki and Ching-Jui Hsiao,
proposed a new algorithm called CHARM, an efficient
algorithm for mining all frequent closed itemsets [9]. It
also uses a technique called diffsets to reduce the memory
footprint of intermediate computations and scalable in the
number of transactions. In 2003, William Cheung and
Osmar R. Zaïane, proposed a novel data structure called
CATS Tree. CATS Tree extends the idea of FPTree to
improve storage compression and allow frequent pattern
mining without generation of candidate itemsets and
allow mining with a single pass over the database as well
as efficient insertion or deletion of transactions at any
time. Later, Zaki in 2001, proposed SPADE, is an
algorithm proposed to find frequent sequences using
efficient lattice search techniques and simple joins [30].
Another approach called MEMory Indexing for
Sequential Pattern mining (MEMISP) is introduced by
Lin and Lee in 2002 [31]. In 2003, Jianyong Wang,
Jiawei Han and Jian Pei, proposed a winning algorithm
CLOSET+. It integrates the advantages of the previously
proposed effective strategies. It is better than existing
mining algorithms including CLOSET, CHARM and OP,
in terms of runtime, memory usage and scalability [10].
Later Guimei Liu Hongjun Lu Wenwu Lou and Jeffrey Xu
Yu, proposed a disk-based data structure, CFP-tree
(Condensed Frequent Pattern Tree), for organizing
frequent patterns discovered from transactional databases
and also developed algorithms to efficiently support two
important types of queries, namely queries with
minimum support constraints and queries with item
constraints, against the stored patterns, as these two types
of queries are basic building blocks for complex frequent
pattern related mining tasks[11].
In 2005, Jianyong Wang, Jiawei Han, Ying Lu, and
Petre Tzvetkov, proposed a mining task: mining top-k
frequent closed itemsets of length no less than min_l,
where k is the desired number of frequent closed itemsets
to be mined, and min_l is the minimal length of each
itemset and developed an efficient algorithm, called TFP,
is developed for mining such itemsets without
min_support [12]. TFP has high performance and linear
scalability in terms of the database size. Later Huiping
Cao, Nikos Mamoulis, and David W. Cheung, modeled
the problem of mining sequential patterns from spatiotemporal data by considering both spatial and temporal
information [13]. In 2007, Huiping Cao, Nikos
Mamoulis and David W. Cheung, solved the problem of
mining periodic patterns in spatiotemporal data and
proposed an effective and efficient algorithm called
STPMine1, for retrieving maximal periodic patterns [32].
In 2010, Zhenhui Li Bolin Ding, Jiawei Han, Roland
Kays and Peter Nye, addressed the problem of mining
periodic behaviors for moving objects and solved with a
two stage algorithm, Periodica [14].
Later, Juyoung Kang and Hwan-Seung Yong, solved
the problem of mining spatio-temporal patterns from
trajectory data. This method first finds meaningful
spatio-temporal regions and extracts frequent spatiotemporal patterns based on a prefix-projection approach
from the sequences of these regions [15]. Later, Nizar R.
Mabroukeh and C. I. Ezeife, proposed a taxonomy of
Sequential Pattern Mining Algorithms, which is
represented in Table1 [16]. Later, Zhenhui Li, Ming Ji,
Jae-Gil Lee, Lu-An Tang, Yintao Yu, Jiawei Han and
Roland Kays, proposed a system called MoveMine, is
designed for sophisticated moving object data mining by
integrating several attractive functions including moving
object pattern mining and trajectory mining [18], [19]. In
2012, P.Mohan and S.Shekhar proposed the cascading
spatiotemporal pattern (CSTP) discovery process finds
partially ordered subsets of these event-types whose
instances are located together and occur serially for a
given collection of Boolean spatiotemporal (ST) eventtypes [24]. Discovering CSTPs from ST data sets is
important for application domains such as public safety
(e.g., identifying crime attractors and generators) and
natural disaster planning, (e.g., preparing for hurricanes).
The proposed model is mentioned section 4.
Table1
Tabular Taxonomy of Sequential-Pattern Mining Algorithms
IV. PROPOSED MODEL
The following are the big data big challenges in the
future era:
i) Data models for managing big data
ii) Big Data Testing – ex. generation of optimal test
data
iii) Real-time streaming data analytics
iv) Scalable analytics on large data sets
v) Systems architecture for big data management
vi) Main memory data management techniques
268
International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 1, January 2014)
vii) Energy-efficient data processing
viii) Energy-aware Resource Management
ix) Benchmarking big data systems
x) Security and Privacy of Big Data
xi) Failover and reliability for big data systems
Fig. 2. Data Stream Processing Engine
V. EXPERIMENTAL RESULTS
We performed all experiments on a 2.2GHz Intel Core
2 Duo processor PC machine with 3GB main memory
Windows XP OS and all programs are implemented in
Java. The results from figure 3(a) to 3(f) show that user
input screen, prefix tree generation, frequent item sets
and generated CFIs & MFIs respectively.
Fig. 1. Architecture for Big Data Pattern Mining
Nowadays it is Impossible for current databases,
technologies and architectures to store and manage huge
data. The above model proposes a solution by analyzing
existing pattern mining algorithms which are mentioned
in Table.1. In 2009, Ho Jin Woo and Won Suk Lee,
proposed estMax algorithm to find closed or maximal
frequent item sets over online transactional data streams
is not easy due to the requirements of a data stream [20].
This method maintains the set of frequent item sets by a
prefix tree and extracts all MFIs without any additional
superset/subset checking mechanism. The Performance
of the estMax method is comparatively analyzed by a
series of experiments to identify its various
characteristics. It has faster processing time and better
performance than estDec algorithm. The main problem
identified in stream data is the processing activity which
is mentioned in the below figure. 2.
Fig. 3(a) Table with list of transactions
Fig. 3(b) List of Transactions in a table
269
International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 1, January 2014)
Fig. 3(c) User Supply Support & Confidence Values
Fig. 3(f) Generated CFIs and MFIs
The future work from this method is to improve the
performance of the proposed method estMax, in terms of
memory usage. To achieve this, a more compact frame of
a prefix tree should be devised. In 2007, Huiping Cao,
Nikos Mamoulis and David W. Cheung, solved the
problem of mining periodic patterns in spatiotemporal
data and proposed an effective and efficient algorithm
called STPMine1, for retrieving maximal periodic
patterns [32]. We implemented the same algorithm with
few modifications in the parameter settings. The
language used was Java and the experiments were
performed on a Pentium IV 700MHz workstation with 2
GB of memory, running Windows XP Operating System.
The future work from this algorithm include the
automatic discovery of the period ‘T’ related to frequent
periodic patterns and the discovery of patterns with
distorted period lengths. The results represented by figure
4(a) to 4(f), show that Data preparation and generated
periodic patterns from a large mobile object movement’s
data.
Fig. 3(d) Generated Prefix Tree
Fig. 3(e) Generated Frequent Item sets
Fig. 4(a) Data Preparation stage
270
International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 1, January 2014)
Fig. 4(b) Data set of objects movements
Fig. 4(e) Region Sequences within Periodicity
Fig. 4(c) Generated Region Sequences
Fig. 4(f) Region-Pattern sequences
In 2010, Jinlin Chen, proposed a novel data structure,
UpDown Directed Acyclic Graph (UDDAG), is invented
for efficient sequential pattern mining. UDDAG allows
bidirectional pattern growth along both ends of detected
patterns. It often outperforms PrefixSpan by almost one
order of magnitude in scalability tests. UDDAG is also
considerably faster than Spade and LapinSpam. We
performed all the experiments on a Windows XP
Operating System with 2.2 GHz Intel Core 2 Duo
Processor PC Machine and 3 GB memory. We
implemented in Java and compared with different
algorithms such as PrefixSpan, Spade, and LapinSpam,
which were all implemented in C++ by previous authors.
The results represented by figure 5(a) to 5(c), show that
generated sequential patterns by using proposed
algorithm called UDDAG with few modifications.
Fig. 4(d) Periodicity Detection (Customized)
271
International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 1, January 2014)
One major feature of UDDAG is that it supports
efficient pruning of invalid candidates. In the future,
there is a requirement of storage of UDDAG in an
efficient way. The future topics include mining with
constraints, closed and maximal pattern mining,
approximate pattern mining, and domain-specific pattern
mining, etc. whenever large searching spaces are
involved and pruning of searching spaces is necessary. In
2011, Faraz Rasheed, Mohammed Alshalalfa, and Reda
Alhajj, proposed an algorithm called STNR,which can
detect symbol, sequence (partial), and segment (full
cycle) periodicity in time series [22]. The algorithm uses
suffix tree as the underlying data structure. The algorithm
is noise resilient; it has been successfully demonstrated to
work with replacement, insertion, deletion, or a mixture
of these types of noise.
We performed all the experiments on a Windows 7
Operating System with 2.3 GHz Intel Core 2 Duo
Processor PC Machine and 3 GB memory. The results
represented by figure 6 (a) to 6(d), show that finding the
whole time series and in a subsection of it to effectively
handle different types of noise. We also implemented
pruning strategy to remove redundant patterns for a given
time-series Database. The conducted study demonstrates
the applicability and effectiveness of the proposed
algorithm; it is generally more time-efficient and noiseresilient than existing algorithms. The future work
includes online construction of suffix tree. This type of
approach is used in sensor networks, ubiquitous
computing, and DNA sequence mining.
Fig. 5(a) User will supply Min_Sup threshold value
Fig. 5(b) Generated Sequential Patterns
Fig. 6(a) Full Series sequence generation
Fig. 5 (c) Sequential Patterns by using UDDAG
272
International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 1, January 2014)
We implemented sanitization algorithm by using java
programming language on a Pentium 2.77GHz processor
with 40 GB Hard disk, 3 GB RAM Windows Operating
System with the back end, MySQL Database. This
algorithm finds spatiotemporal trajectories which require
hiding of trajectories that are unexploitable from
available data.There are different ways to hide sensitive
trajectory data considered as a future work from this
approach.
Recently in 2012, Mr. Pradeep Mohan et al. proposed
QSTP discovery process which finds the partially ordered
subsets of the Boolean Spatio-Temporal event-types
whose instances are located together & occur serially
[24].The discovered patterns are represented using
Directed Acyclic Graph (DAG). Originally, this
algorithm was implemented in MATLAB 7.9 and we
implemented this approach with modifications by using
PHP with the back end MySQL and performed
experiments on windows Operating System with 3 GB
Memory. The results from figure 7(a) to 7(c) show that
how to get cascading spatiotemporal patterns called
CSTPM. The future work includes how to perform
expensive significance tests by randomly permuting ST
instances and accounting for multiple testing, which is
used in weather forecasting.
Fig. 6(b) periodicity within a subsection
Fig. 6(c) Pruning strategy on Full Series patterns
Fig. 7(a) QSTPs used in Crime Detection events
Fig. 6(d) Pruning strategy on subsection series
In 2010, Osman Abul, Francesco Bonchi, and Fosca
Giannotti, proposed the problem of hiding sequential
patterns and show it’s NP-hardness [23].
Fig. 7(b) Generated CSTPs
273
International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 1, January 2014)
[6]
[7]
Fig. 7(c) Resulted CSTPs
[8]
VI. CONCLUSIONS AND FUTURE WORK
In this paper, we proposed a new framework useful in
Big Data pattern mining including the observations of
many pattern mining algorithms with different scenarios.
The future work contains how to mine frequent patterns
and closed frequent patterns, sequential patterns, Periodic
patterns with different parameters in Big Data field. The
traditional algorithms and techniques may not be suitable
for big data due to large, heterogeneous and stream data
needs to be processed with in elapsed time [40]. Finally,
there is a scope to develop new algorithms and methods
works on MapReduce programming model to perform
pattern mining, an active area of research in Big Data
domain.
[9]
[10]
[11]
[12]
Acknowledgements
[13]
The authors appreciate the efforts of UG Students, Sai
Vamsi.V, M.Sandeep Reddy, M.Girish Kumar Reddy,
P.Keerthi and Sai Avinash and Research Scholar,
K.Suresh including the comments from the Professor
Dr.A.Rama Mohan Reddy, Department of CSE,
S.V.University College of Engineering; their feedback
and suggestions improved the quality of the research
presented here. The views expressed here are those of the
authors and the authors thank the handling associate
editors and the anonymous reviewers for their
constructive comments.
[14]
[15]
[16]
REFERENCES
[1]
[2]
[3]
[4]
[5]
[17]
Rakesh Agrawal Ramakrishnan Srikant. Fast Algorithms for
Mining Association Rules. Proceedings of the 20th VLDB
Conference Santiago, Chile, 1994.
R. Srikant, R. Agrawal: "Mining Sequential Patterns:
Generalizations and Performance Improvements", Proc. of the
Fifth Int'l Conference on Extending Database Technology
(EDBT), Avignon, France, March 1996. Expanded version
available as IBM Research Report RJ 9994, December 1995.
Roberto J. Bayardo Jr. Efficiently Mining Long Patterns from
Databases. Proc. of the 1998 ACM-SIGMOD Int’l Conf. on
Management of Data, 85- 93.
J. Pei, J. Han, and R. Mao. "CLOSET: An efficient algorithm for
mining frequent closed itemsets". Proceedings of the 2000 ACM
SIGMOD Workshop on Research Issues in Data Mining and
Knowledge Discovery, Dallas, TX, May, 2000.
Yves Bastide, Rafik Taouil, Nicolas Pasquier, Gerd Stumme.
Mining Frequent Patterns with Counting Inference. SIGKDD
Explorations. ACM SIGKDD, December 2000. Volume 2, Issue 2
- page 66-75.
[18]
[19]
[20]
[21]
[22]
274
J. Han, J. Pei, Y. Yin and R. Mao, "Mining Frequent Patterns
without Candidate Generation: A Frequent-Pattern Tree
Approach", Data Mining and Knowledge Discovery, an
International Journal, Volume 8, Issue 1, pages 53-87, January
2004, Kluwer Academic Publishers.
Doug Burdick, Manuel Calimlim, Johannes Gehrke. MAFIA: A
Maximal Frequent Itemset Algorithm for Transactional
Databases. ICDE, 2001.
Ilias Tsoukatos and Dimitrios Gunopulos. Efficient Mining of
Spatiotemporal Patterns. C.S. Jensen et al. (Eds.):SSTD 2001,
LNCS 2121, pp. 425−442, 2001. Springer-Verlag Berlin
Heidelberg 2001.
Mohammed J. Zaki and Ching-Jui Hsiao, CHARM: An Efficient
Algorithm for Closed Itemset Mining. 2nd {SIAM} International
Conference on Data Mining. April 2002.
J. Wang, J. Han, and J. Pei, "CLOSET+: Searching for the Best
Strategies for Mining Frequent Closed Itemsets", Proc. 2003
ACM SIGKDD Int. Conf. on Knowledge Discovery and Data
Mining (KDD'03), Washington, D.C., August, 2003.
Guimei Liu Hongjun Lu Wenwu Lou and Jeffrey Xu Yu. On
Computing, Storing and Querying Frequent Patterns.
SIGKDD’03, August24-27, 2003, Washington, DC, USA. 2003
ACM 1-58113-737-0/03/0008 ...$5.00.
J. Wang, J. Han, Y. Lu, and P. Tzvetkov, "TFP: An Efficient
Algorithm for Mining Top-K Frequent Closed Itemsets", IEEE
Transactions on Knowledge and Data Engineering, 17(5):652664, 2005.
H. Cao, N. Mamoulis, and D. W. Cheung, "Mining Frequent
Spatio-temporal Sequential Patterns," Proceedings of the 5th IEEE
International Conference on Data Mining (ICDM), pp. 82-89,
Houston, Texas, November 2005.
Zhenhui Li Bolin Ding, Jiawei Han, Roland Kays and Peter Nye.
Mining Periodic Behaviors for Moving Objects. KDD’10, July
25–28, 2010, Washington, DC, USA. 2010 ACM 978-1-45030055-1/10/07 ...$10.00.
Juyoung Kang and Hwan-Seung Yong. Mining Spatio-Temporal
Patterns in Trajectory Data. Journal of Information Processing
Systems, Vol.6,
No.4, December 2010 DOI :
10.3745/JIPS.2010.6.4.521.
Mabroukeh, N. R. and Ezeife, C. I. 2010. A Taxonomy of
sequential pattern mining algorithms. ACM Comput. Surv. 43, 1,
Article 3, November 2010, 41 pages.
Rakesh Agrawal Sakti Ghosh Tomasz Imielinski, Bala Iyer, Arun
Swami. An Interval Classier for Database Mining Applications.
Proceedings of the 18th VLDB Conference Vancouver, British
Columbia, Canada 1992.
Zhenhui Li, Ming Ji, Jae-Gil Lee, Lu-An Tang, Yintao Yu, Jiawei
Han and Roland Kays, MoveMine: Mining Moving Object
Databases. SIGMOD’10, June 6–11, 2010, Indianapolis, Indiana,
USA. 2010 ACM 978-1-4503-0032-2/10/06 ...$10.00.
Qiankun Zhao and Sourav S. Bhowmick. Sequential Pattern
Mining: A Survey. Technical Report, CAIS, Nanyang
Technological University, Singapore, No. 2003118 , 2003.
Ho Jin Woo and Won Suk Lee. estMax: Tracing Maximal
Frequent Item Sets Instantly over Online Transactional Data
Streams. IEEE
Transactions on Knowledge and Data
Engineering, VOL. 21, NO. 10, OCTOBER 2009.
Jinlin Chen. An UpDown Directed Acyclic Graph Approach for
Sequential Pattern Mining. IEEE Transactions on Knowledge and
Data Engineering, VOL. 22, NO. 7, JULY 2010.
Faraz Rasheed, Mohammed Alshalalfa, and Reda Alhajj. Efficient
Periodicity Mining in Time Series Databases Using Suffix Trees.
IEEE Transactions on Knowledge and Data Engineering, VOL.
23, NO. 1, JANUARY 2011.
International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 1, January 2014)
[23] Osman Abul, Francesco Bonchi, and Fosca Giannotti, Hiding
Sequential and Spatiotemporal Patterns. IEEE Transactions on
Knowledge and Data Engineering, VOL. 22, NO. 12,
DECEMBER 2010.
[24] Pradeep Mohan, Shashi Shekhar, James A. Shine, and James P.
Rogers. Cascading Spatio-Temporal Pattern Discovery. IEEE
Transactions on Knowledge and Data Engineering, VOL. 24,
NO. 11, NOVEMBER 2012.
[25] Xixian Han, Jianzhong Li, Donghua Yang, and Jinbao Wang.
Efficient Skyline Computation on Big Data. IEEE Transactions on
Knowledge and Data Engineering, VOL. 25, NO. 11, November
2013.
[26] Manish Bhide and Krithi Ramamritham. Category-Based
Infidelity Bounded Queries over Unstructured Data Streams.
IEEE Transactions on Knowledge and Data Engineering, VoL.
25, No. 11, November 2013.
[27] J.Han, M.Kamber and J.Pei. Data Mining: Concepts and
Techniques. Morgan Kauffman Series. 3rd Edition, 2012.
[28] Wei Fan, Albert Bifet. Mining Big Data: Current Status, and
Forecast to the Future. SIGKDD Explorations Volume 14, Issue 2,
Pno.s 1-5.
[29] M.N.Garofalakis, Rajeev Rastogi and Kyuseok Shim. SPIRIT:
Sequential Pattern Mining with Regular Expression Constraints.
Proceedings of VLDB '99 Proceedings of the 25th International
Conference on Very Large Databases, Page Nos. 223-234.
[30] Mohammed J. Zaki, SPADE: An Efficient Algorithm for Mining
Frequent Sequences. Machine Learning Journal, 42(1/2):31-60,
2001.
[31] Ming-Yen Lin and Suh-Yin Lee. Fast Discovery of Sequential
Patterns through Memory Indexing and Database Partitioning.
Springer Berlin Heidelberg, 2002.
[32] Huiping Cao, Nikos Mamoulis, David Wai-lok Cheung.
Discovery of Periodic Patterns in Spatiotemporal Sequences,
IEEE TKDE, Vol. 19, No. 4, April 2007 pp. 453-467.
[33] O.Obulesu, A. Rama Mohan Reddy and K. Suresh. Finding
Maximal Periodic Patterns and Pruning Strategy in
Spatiotemporal Databases. IJARCSSE, Vol.2, Issue No.4, April
2012. ISSN: 2277 128X Page No.s 423-426.
275