16 (IJCNS) International Journal of Computer and Network Security,

Vol. 2, No. 8, August 2010

An Adoptive Algorithm for Mining Time-Interval

Sequential Patterns
Hao-En Chueh1 and Yo-Hsien Lin2
1
Department of Information Management, Yuanpei University,
No.306, Yuanpei Street, Hsinchu 30015, Taiwan, R.O.C.
hechueh@mail.ypu.edu.tw
2
Department of Information Management, Yuanpei University,
No.306, Yuanpei Street, Hsinchu 30015, Taiwan, R.O.C.
yohsien@mail.ypu.edu.tw
customers at the right time. Therefore, recently, some
Abstract: An adoptive algorithm for mining time-interval
sequential patterns is presented in this paper. A time-interval researches start to propose algorithms for discovering the
sequential pattern is a sequential pattern with time-intervals sequential patterns with time-intervals between successive
between successive itemsets. Most proposed algorithms use some itemsets, this kind of pattern is called time-interval
predefined non-overlap time partitions to find the time-intervals sequential pattern [2].
between successive itemsets, but a predefined set of non-overlap
time partitions cannot be suitable for every pair of successive To discover the time-interval sequential patterns, many
itemsets. Therefore, in this paper, the clustering analysis is used researches adopt some predefine non-overlap time
first to generate the suitable time-intervals for frequent partitions, and assume that the time-intervals between
occurring pairs of successive itemsets. Next, the generated time- successive itemsets of the frequently sequential patterns can
intervals are used to extend the typical sequential patterns
fit into one of the predefined time partitions.
mining algorithms to discover the time-interval sequential
patterns. Finally, an operator to obtain the time-interval of a However, a predefined set of non-overlap time partitions
subsequence of a time-interval sequential pattern is also cannot be suitable for every pair of successive itemsets.
presented.
Therefore, generating the suitable time partitions for every
Keywords: Adoptive algorithm, Time-Interval, Sequential pair of successive itemsets directly from the real sequence
Pattern, Clustering Analysis datasets is more reasonable. Accordingly, in this paper, we
present an adoptive algorithm to discover the time-interval
1. Introduction sequential patterns without using predefined time partitions.
Data mining is usually defined as the procedure of This algorithm uses clustering analysis to automatically
discovering hidden, useful, previously unknown information obtain suitable time-intervals between frequent occurring
from large databases. The common data mining techniques pairs of successive itemsets, and then uses these time-
include classification, clustering analysis, association rules intervals to extend typical sequential patterns mining
mining, sequential patterns mining and so on. Sequential algorithms to discover the time-interval sequential patterns.
patterns mining introduced by Agrawal and Srikant (1995) The rest of this paper is organized as follows. Some
is the task of finding frequently occurring patterns related to researches related to time-interval sequential patterns are
time or other sequences from a given sequence database [1]. reviewed in section 2. The proposed time-interval sequential
It is widely used in the field of retail business to assist in patterns mining algorithm is presented in section 3. A
making various marketing decisions [3, 5, 7]. An example example is displayed in section 4. The conclusion is given in
of a sequential pattern is “A customer who bought a digital section 5.
camera will buy a battery and a memory card later”.
Up to now, many algorithms have been proposed [1, 4, 6,
9] for mining sequential patterns, however, most of these 2. Time-Interval Sequential Patterns
algorithms only focus on the order of the itemsets, but Sequential patterns mining is defined as the task of
ignore the time-intervals between itemsets. In business field, discovering frequently occurring ordered patterns from the
actually, a sequential pattern which includes the time- given sequence database. A sequence is an ordered list of
intervals between successive itemsets is more valuable than itemsets. Let I = { i1 , i2 ,......,im } be a set of items,
a sequential pattern without any time information. An
S =< s1 , s 2 , ......, s k > is a sequence, where s i ⊆ I is called an
example of a sequential pattern with time intervals between
successive itemsets is “A customer who bought a digital itemset. Length of a sequence means the number of itemsets
camera will return to buy a battery and a memory card in the sequence, and a sequence contains k itemsets is called
within one week”. Clearly, the time-intervals between a k-sequence. The support of a sequence S is denoted by
itemsets can offer the retail business more useful supp (S ) and means the percentage of total number of
information to sell the appropriate products to their records containing sequence S in the sequence database. If
 (IJCNS) International Journal of Computer and Network Security, 17
supp (S ) is greater than or equal to a predefined threshold,
called minimal support, than sequence S is regarded as a
frequent sequence and called a sequential pattern.
Many algorithms have been proposed to discover
sequential pattern [1, 4, 6, 9], and most algorithms only
focus on the frequently occurring order of the itemsets, but
ignore the time-intervals between itemsets. The time-
intervals between successive itemsets, in fact, can offer
useful information for business to sell the appropriate
products to their customers at the right time. Due to the
value of the time-intervals between successive itemsets,
many algorithms for mining various sequential patterns with
time-intervals between successive itemsets have been
proposed [2, 8, 10, 11].
Srikant et al. [10] utilize three predefined restrictions, the
maximum interval (max − interval), the minimum interval
(min − interval), and the time window size ( window − size)
to find sequential patterns related to time-intervals. The
discovered sequential pattern is like (( A, B), (C , D)), where
( A, B) and (C , D) are two subsequences of (( A, B), (C , D)).
The max − interval and min − interval are respectively
used to indicate the maximal time-interval and the minimal
time- interval within subsequence. The window − size is
used to indicate the time-interval among subsequences.
Assume that the max − interval is set to 10 hours, the
min − interval is set to 3 hours, and the window − size is
set to 24 hours, then the time-interval between A and B
lies in [ 2, 10] , the time- interval between C and D also
lies in [ 2, 10] , and the time- interval
between ( A, B) and (C , D) lies in [ 0, 24] .
Mannila et al. [8] use a predefined window width (win)
to find frequent episodes in sequences of events, and the
discovered episode is like ( A, B, C ). Assume that the win is
set to 3 days, then the episode ( A, B, C ) means that, in 3
days, A occurs first, B follows, and C happens finally.
Wu et al. [11] also utilize a window (d ) to find the
sequential pattern likes ( A, B, C ), such that, in the sequential
pattern ( A, B, C ), the time-interval between adjacent events
is within d . Assume d is set to 5 hours, then the
discovered pattern ( A, B, C ) means that A occurs first, B
follows, and C happens finally; the time-interval between
A and B , and the time-interval between B and C are
both within 5 hours.
Chen et al. [2] use a predefined set of non-overlap time
partitions to discover potential time-interval sequential
patterns, and the discovered pattern is like ( A, I 0 , B, I 2 , C ),
where I 0 , I 2 belong to the non-overlap set of time partitions.
Assume that, I 0 denotes the time-interval t satisfying
0 ≤ t ≤ 1 day; I 2 denotes the time interval t satisfying
3 < t ≤ 7 days, and then the pattern ( A, I 0 , B, I 2 , C ) means
Vol. 2, No. 8, August 2010
that A, B and C happen in this order, and the time-interval
between A and B is within 1 day , and the time-interval
between B and C lies between 3 days and 7 days.
These proposed researches can discover the sequential
patterns with the time-intervals between successive itemsets
by using a or some predefined time partitions, but the
sequential patterns with time-intervals between successive
itemsets lie outside these used time ranges cannot be found
yet. To solve this problem, therefore, an adoptive algorithm
for mining time-interval sequential patterns without using
any predefined time partitions is presented in this work. The
main concept of this proposed algorithm is to generate the
suitable time-intervals directly from the real sequence
dataset. The algorithm first adopts clustering analysis to
automatically generate the suitable time-intervals for
frequent occurring pairs of successive itemsets, and then
uses these time-intervals to extend typical algorithms to
discover sequential patterns with time-ntervals between
successive itemsets. Details of the proposed algorithm are
introduced in the next section.
3. Adoptive Time-Interval Sequential Patterns
Mining Algorithm
The proposed algorithm for mining time-interval sequential
patterns is introduced as follows. First, some notations are
defined in advance.
I = { i1 , i 2 ,......,i m } : The set of items.
Si =< s1 , s 2 ,......, s n > : A sequence, where each s k ⊆ I .
D = { S1 , S2 ,......, Sk } : The sequences dataset.
supp ( Si ) : The support of the sequence Si .
min − supp : The minimal support threshold.
CSk : The candidate set of frequent k-sequences.
FSk : The set of frequent k-sequences.
CTISk : The candidate set of frequent time-interval k-
sequences.
FTIS k : The set of frequent time-interval k-sequences.
3.1 The proposed algorithm
Step 1: Produce FS1 , the set of frequent 1-sequences.
Each items s i ∈ I is as a candidate frequent 1-sequence. A
candidate frequent 1-sequence whose support is greater than
or equal to min − supp is a frequent 1- sequence, and FS1
denotes the set of all frequent 1-sequences.
Step 2: Produce CS2 , the candidate set of frequent 2-
sequences. From any two frequent 1- sequences of FS1 , say
s1 and s2 , where s1 , s2 ∈ FS1 and s1 ≠ s 2 , generate 2
candidate frequent 2-sequences belong to CS2 ,
say < s1 , s 2 > and < s2 , s1 > .
18 (IJCNS) International Journal of Computer and Network Security,
Vol. 2, No. 8, August 2010

Step 3: Produce FS2 , the set of frequent 2-sequences. A
candidate frequent 2-sequence whose support is greater than
or equal to min − supp is a frequent 2-sequence, and the set
of all frequent 2-sequences is FS2 .
Step 4: Find the frequent time-intervals for each 2-
sequence of FS2 . For any frequent 2-sequence of FS2 ,
say < s p , s q >, all the time-intervals between s p and s q
appear in D and are listed in increasing order, then the
following clustering analysis based steps, Step 4(a), Step
4(b) and Step 4(c) are used to obtain the frequent time-
intervals.
Step 4(a): Let T (1, z) = [t1 , t 2 ,L, t z ] is the increasingly
ordered list of the time-intervals of
< s p , sq > . Let T < s p , s q > = {T (1, z) } be the set of time-
time-interval (k-1)-sequences S1 and S2 ,
where S1 =< s1,1 , T1,1 , s1, 2 ,L, s1,k − 2 ,
T1,k − 2 , s1, k −1 >, S2 =< s 2 ,1 , T2,1 , s 2 , 2 ,L, s 2 ,k − 2 , T2 ,k − 2 , s 2 ,k −1 >∈ FTISk −1 ;
s1, 2 = s 2 ,1 , s1, 3 = s 2 , 2 ,L, s1,k −1 = s 2, k − 2 ; T1, 2 = T2 ,1 , T1, 3 = T2 , 2 ,L, T1,k − 2
= T2 ,k −3 , then we can generate a candidate time-interval k-
sequence S12 =< s1,1 , T1,1 , s1, 2 ,L, s1,k − 2 , T1, k − 2 , s1,k −1 , T2 ,k − 2 , s 2, k −1 > .
Step 7: Produce FTISk , k ≥ 3 , the set of frequent time-
interval k-sequences. A candidate time-interval k-sequence
whose support is greater than or equal to min − supp is a
frequent time-interval k-sequence, and the set of frequent
time- interval k-sequences is FTISk .
Step 8: Repeat Step 6 and Step 7, until no next CTIS k can
be generated.

intervals of < s p , sq > . The first step is to find the maximal 3.2 Time-intervals of a subsequence
difference between two adjacent time-intervals of In this subsection, an operator to obtain the time-interval of
T < s p , s q >, and then divide T < s p , sq > into 2 subsets a subsequence of a frequent time-interval sequential pattern
according to the maximal difference. Assume that the is introduced. Let S =< s1 , T1, 2 , s 2 ,L, sk −1 , Tk −1, k , s k > is a
difference between ti and t i+1 is maximal, then T (1, z) is frequent time-interval k-sequence, and
divided into T (1, i ) and T (i + 1, z), where T (1, i ) = [t1 ,L, t i ], S ' =< si , Ti ,i +1 , s i=1 ,L, si + j −1 , Ti + j −1,i+ j , s i+ j > is a subsequence
T (i + 1, z) = [t i +1 ,L, t z ]. of S. Assume that Ti ,i+1 =
[a i , bi ], Ti +1,i + 2 = [a i+1 , bi +1 ],L, Ti+ j −1,i+ j = [a i+ j −1 , bi + j −1 ], and then
Step 4(b): Calculate the support of < s p , sq > that
the time-interval between si and s i + j is equal
respectively includes each time- interval set. If the support
to [a i + a i+1 + L + a i + j −1 , bi + bi +1 + Lbi+ j −1 ].
of < s p , sq > that includes time-intervals T (1, i ) is greater
By using the above steps, a simple example is displayed
than or equal to min − supp, then T (1, i ) is a frequent time-
in the next section.
interval of < s p , sq > , and then T (1, i ) is reserved, otherwise
T (1, i ) is deleted. Similarly, if the support of < s p , sq > that
4. Example
includes time-intervals T (i + 1, z) is greater than or equal to In this section, we use the example sequence database
shown as in Table 1 to discover the time-interval sequential
min − supp, then T (i + 1, z) is also a frequent time-interval,
patterns. In Table 1, Id denotes the record number of a
and T (i + 1, z) is reserved, otherwise T (i + 1, z) is deleted. sequence, and each sequence is represented
The reserved subsets of time-intervals next replaces the as < (s1 , t1 ), (s 2 , t 2 ),L, ( s n , t n ) >, where s i denotes an itemset,
original set of time-intervals T < s p , sq > . If no subset is and t i denotes the time stamp that s i occurs; here, the
reserved, then the original set of time-intervals is called min − supp is set to 0.3.
non- dividable. If all differences between two adjacent time-
intervals in the original set of time-intervals are equal, then Table 1: A sample sequence database
the original set of time-intervals is called as non-dividable Id Sequence
as well. 01 ( s 5 ,8), ( s 4 ,15), ( s 6 ,20)

Step 4(c): Repeat Step 4(a) and Step 4(b), until all subsets
of time-intervals in T < s p , sq > are non-dividable.
Step 5: Produce FTIS 2 , the set of frequent time-interval
2- sequences. Each 2-sequence of FS2 is extended by all its
frequent time-intervals to generate FTIS 2 . If T < s p , s q >=
{T 1 , T 2 ,L, T R } is the set of frequent time-intervals of
< s p , s q >, then T < s p , T i , s q >, i = 1L R, is a frequent time-
interval 2- sequence.
Step 6: Produce CTISk , k ≥ 3 , the candidate set of
frequent time-interval k-sequences. For any two frequent
02 ( s1 ,2), ( s 3 ,7), ( s 2 ,11), ( s 6 ,18)
03 ( s 2 ,3), ( s1 ,4), ( s 3 ,7), ( s 6 ,16), ( s 7 ,19)
04 ( s1 ,2), ( s 2 ,8), ( s 6 ,10), ( s 7 ,15)
05 ( s 5 ,4), ( s 6 ,16), ( s1 ,20), ( s 3 ,24)
06 ( s 7 ,7), ( s1 ,13), ( s 5 ,18), ( s 2 ,25), (s6,28)
07 ( s 5 ,4), ( s1 ,8), ( s 3 ,12), ( s 6 ,16), ( s 7 ,20)
( s1 ,3), ( s 5 ,6), ( s 2 ,9), ( s 4 ,18), ( s 6 ,21)
08
09 ( s 2 ,5), ( s1 ,10), ( s 3 ,15), ( s 6 ,20), ( s 7 ,25)
10 ( s 6 ,2), ( s 7 ,8), ( s 5 ,12), ( s 2 ,17)
First, we need to calculate the supports of all itemsets to
produce FS1 . Supports of all itemsets are shown in Table 2.
 (IJCNS) International Journal of Computer and Network Security, 19
Vol. 2, No. 8, August 2010

Here, we can obtain FS1 = {s1 , s 2 , s 3 , s 5 , s6 , s7 }. < s3 , s7 > T < s 3 , s 7 >={8, 10, 12}
< s5 , s2 > T < s 5 , s 2 >={3, 5, 7}
Table 2: Supports of itemsets < s5 , s6 > T < s 5 , s 6 >={10, 12, 15}
itemsets support
< s6 , s7 > T < s 6 , s 7 >={3, 4, 5, 6 }
s1 0.8
s2 0.7 According to the step 4 described in the section 3, the set
s3 of all suitable time-intervals for each sequences of FS2 are
0.5
s4 obtained as in Table 5.
0.2
s5 0.6 Next, each 2-sequence of FS2 is extended by all its
s6 1 suitable time-intervals to form FTIS 2 .
s7 0.6 1 1 1
FTIS 2 ={< s1 , T , s 2 >, 1, 2 < s1 , T , s 3 >,
1, 3 < s1 , T , s 6 >,
1, 6

Next, CS2 is generated by jointing FS1 × FS1 ; Supports of < s1 , T1,26 , s 6 >, < s1 , T11, 7 , s 7 >, < s 2 , T21, 6 , s 6 >, < s 2 , T22, 6 , s 6 >,
the sequences in CS2 are calculated and shown in Table 3. < s 2 , T21, 7 , s 7 >, < s 3 , T31, 6 , s 6 >, < s 3 , T31, 7 , s 7 >, < s 5 , T51, 2 , s 2 >,
Therefore, we obtain FS3 = {< s1 , s 2 >, < s1 , s3 >, < s1 , s 6 >,
< s 5 , T51, 6 , s 6 >, < s 6 , T61, 7 , s 7 >}.
< s1 , s 7 >, < s 2 , s6 >, < s 2 , s 7 >, < s3 , s 6 >, < s3 , s7 >, < s5 , s 2 >,
< s5 , s 6 >, < s 6 , s 7 >}.
Table 5: Suitable time-intervals of the sequences in FS2
For each frequent 2-sequence of FS2 , all its time FS2 time-intervals
intervals are recorded and listed in increasing order (Table < s1 , s 2 > T < s1 , s 2 >={ T11, 2 = [6, 12] }
4).
< s1 , s 3 > T < s1 , s 3 >={ T11, 3 = [3, 5] }
< s1 , s 6 > T < s1 , s 6 >={ T11, 6 = [8,12], T1,26 = [16,18] }
Table 3: Supports of sequences in CS2
< s1 , s 7 > T < s1 , s 7 >={ T11,7 = [12,15] }
CS2 support CS2 support
< s2 , s6 > T < s 2 , s 6 >={ T21, 6 = [ 2,7], T22, 6 = [12,15] }
< s1 , s 2 > 0.4 < s 5 , s1 > 0.2
< s2 , s7 > T < s 2 , s 7 >={ T21,7 = [16,20] }
< s1 , s 3 > 0.5 < s5 , s2 > 0.3
< s3 , s6 > T < s 3 , s 6 >={ T31,6 = [4,11] }
< s1 , s 5 > 0.2 < s5 , s3 > 0.2
< s3 , s7 > T < s 3 , s 7 >={ T31,7 = [8,12] }
< s1 , s 6 > 0.7 < s5 , s6 > 0.5
< s5 , s2 > T < s 5 , s 2 >={ T51,2 = [3,7] }
< s1 , s 7 > 0.4 < s5 , s7 > 0.1
< s5 , s6 > T < s 5 , s 6 >={ T51,6 = [10,12] }
< s 2 , s1 > 0.2 < s 6 , s1 > 0.1
< s6 , s7 > T < s 6 , s 7 >={ T61,7 = [3,6] }
< s2 , s3 > 0.2 < s6 , s2 > 0.1
< s2 , s5 > 0.0 < s6 , s3 > 0.1 CTIS3 , the candidate set of frequent time-interval 3-
< s2 , s6 > 0.6 < s6 , s5 > 0.1 sequences is generated by jointing FTIS2 × FTIS 2 . Supports
< s2 , s7 > 0.3 < s6 , s7 > 0.5 of the sequences of CTIS3 are calculated and shown in Table
< s 3 , s1 > 0.0 < s 7 , s1 > 0.1 6. A candidate frequent time-interval 3-sequence whose
support is greater than or equal to min − supp is called as a
< s3 , s2 > 0.1 < s7 , s2 > 0.2
frequent time-interval 3-sequence. Therefore, we can obtain
< s3 , s5 > 0.0 < s7 , s3 > 0.0
the set of all the frequent time-interval 3-sequences,
< s3 , s6 > 0.4 < s7 , s5 > 0.2 FTIS3 = {< s1 ,
< s3 , s7 > 0.3 < s7 , s6 > 0.1
T11, 2 , s 2 , T21, 6 , s 6 >, < s1 , T11, 3 , s3 , T31, 6 , s6 >, < s1 , T11,3 , s3 , T31, 7 , s 7 >,
< s1 , T11,6 , s6 , T61,7 , s 7 >, < s3 , T31, 6 , s 6 , T61, 7 , s7 >}.
Table 4: Time-intervals of the sequences in FS2
FS2 time-intervals
Table 6: Supports of sequences in CTIS3
< s1 , s 2 > T < s1 , s 2 >={6, 9, 12}
CTIS3 support
< s1 , s 3 > T < s1 , s 3 >={3, 4, 5}
< s1 , T , s 2 , T , s 6 >
1
1, 2
1
2,6
0.3
< s1 , s 6 > T < s1 , s 6 >={8, 10, 12, 16, 18}
< s1 , T11,2 , s 2 , T22,6 , s 6 > 0.1
< s1 , s 7 > T < s1 , s 7 >={12, 13, 15}
< s1 , T11,2 , s 2 , T21,7 , s 7 > 0.1
< s2 , s6 > T < s 2 , s 6 >={2, 3, 7, 12, 13, 15}
< s1 , T11,3 , s 3 , T31,6 , s 6 > 0.4
< s2 , s7 > T < s 2 , s 7 >={7, 16, 20}
< s1 , T , s 3 , T , s 7 >
1
1, 3
1
3,7
0.3
< s3 , s6 > T < s 3 , s 6 >={4, 5, 9, 11}
< s1 , T , s 6 , T , s 7 >
1
1, 6
1
6,7
0.4
20
< s 2 , T21,6 , s 6 , T61,7 , s 7 > 0.1
< s2 , T , s6 , T , s7 >
2 1
2,6 6,7
0.2
< s3 , T , s6 , T , s7 >
1
3, 6
1
6,7
0.3
< s5 , T , s2 , T , s7 >
1
5,2
1
2,7
0 [3] M. S. Chen, J. Han, P. S. Yu, “Data mining: An
< s 5 , T51,6 , s 6 , T61,7 , s 7 > 0.1
The candidate set of frequent time-interval 4-sequences,
CTIS4 , is generated by jointing FTIS3 ×FTIS3 . Here, only
one sequence, < s1 , T11,3 , s 3 , T31,6 , s 6 , T61, 7 , s 7 >, is generated.
The support of the sequence < s1 , T11,3 , s 3 , T31,6 , s 6 , T61, 7 , s 7 > is
0.3, thus < s1 , T11,3 , s 3 , T31,6 , s 6 , T61, 7 , s 7 > is also a frequent
time- interval 4-sequences, and
FTIS4 = < s1 , T11,3 , s 3 , T31,6 , s 6 , T61, 7 , s 7 >.
next CTIS5 can be generated, the algorithm stops here. In
addition, the time-interval of any subsequence of
< s1 , T11,3 , s 3 , T31,6 , s 6 , T61, 7 , s 7 > can be obtained by using the
operator introduced in subsection 3.2.
From the above example, we can clearly see that the
suitable time-intervals for every pair of successive itemsets
are different and overlap, therefore, it is more reasonable to
generate the suitable time-intervals directly from the real
sequence data for every pair of successive itemsets when
mining time-interval sequential patterns.
5. Conclusion
In this paper, we present an adoptive algorithm for mining
time-interval sequential patterns. A sequential pattern with
the time-intervals between successive itemsets is more
valuable than a traditional sequential pattern without any
time information. Most proposed algorithms reveal the time-
intervals between itemsets by using some predefined non-
overlap time partitions, but this way, in fact, may not be
suitable for every pair of successive itemsets. To solve this
problem, the proposed algorithm uses clustering analysis to
automatically generate the suitable time-intervals between
frequent occurring pairs of successive itemsets, and then
uses these generated time-intervals to extend typical
algorithms to discover the time-interval sequential patterns
without pre- defining any time partitions. In addition, a
useful operator for computing the time-interval of a
subsequence of a frequent time-interval sequential pattern is
also introduced in this paper. From the result of the
example, we can conclude that because the time-intervals
between successive itemsets are quite different and overlap,
it is more reasonable to generate the suitable time-intervals
directly from the real sequence data when mining time-
interval sequential patterns.
References
[1] R. Agrawal, R. Srikant, “Mining sequential patterns,”
In Proceedings of the International Conference on Data
Engineering, pp. 3-14, 1995.
(IJCNS) International Journal of Computer and Network Security,
Vol. 2, No. 8, August 2010
[2] Y. L. Chen, M. C. Chiang, M. T. Ko, “Discovering
time-interval sequential patterns in sequence
databases,” Expert Systems with Applications, 25(3),
pp. 343-354, 2003.
overview from a database perspective,” IEEE
Transactions on Knowledge and Data Engineering,
8(6), pp. 866-883, 1996.
[4] M. S. Chen, J. S. Park, P. S. Yu, “Efficient data
mining for path traversal patterns,” IEEE Transactions
on Knowledge and Data Engineering, 10(2), pp. 209-
221, 1998.
[5] M. H. Dunham, Data mining, Introductory and
Advanced Topics, Pearson Education Inc., 2003.
we obtain [6] J. Han, G. Dong, Y. Yin, “Efficient mining of partial
periodic patterns in time series database,” In
Because no
Proceedings of the 1999 International Conference on
Data Engineering, pp. 106-115, 1999.
[7] J. Han, M. Kamber, Data mining: Concepts and
Techniques, Academic Press, 2001.
[8] H. Mannila, H. Toivonen, A. Inkeri Verkamo,
“Discovery of frequent episodes in event sequences,”
Data Mining and Knowledge Discovery, 1(3), pp. 259-
289, 1997.
[9] J. Pei, J. Han, H. Pinto, Q, Chen, U. Dayal, M.-C. Hsu,
“PrefixSpan: Mining sequential patterns efficiently by
prefix-projected pattern growth,” In Proceedings of
2001 International Conference on Data Engineering,
pp. 215-224, 2001.
[10] R. Srikant, R. Agrawal, “Mining sequential patterns:
Generalizations and performance improvements,” In
Proceedings of the 5th International Conference on
Extending Database Technology, pp. 3-17, 1996.
[11] P. H. Wu, W. C. Peng, M. S. Chen, “Mining sequential
alarm patterns in a telecommunication database,” In
Proceedings of Workshop on Databases in
Telecommunications (VLDB 2001), pp. 37-51, 2001.
Authors Profile
Hao-En Chueh received the Ph.D. in
Computer Science and Information
Engineering from Tamkang University,
Taiwan, in 2007. He is an Assistant
Professor of Information Management at
Yuanpei University, Hsinchu, Taiwan. His
research interests include data dining,
fuzzy set theory, probability theory,
statistics, database system and its
applications.
Yo-Hsien Lin received the Ph.D. in
information management from the
National YunLin University of Science
and Technology, Taiwan, in 2008. He is
an Assistant Professor of Information
Management at the Yuanpei University,
Hsinchu, Taiwan. His research interests
include bio-inspired systems, neural
networks, evolutionary computation,
evolvable hardware, intelligence system
chip, biocomputing, pattern recognition,
and medical information management.