Sequential Pattern Mining

Outline
What is sequence database and sequential pattern mining Methods for sequential pattern mining Constraint-based sequential pattern mining Periodicity analysis for sequence data

Sequence Databases
A sequence database consists of ordered elements or events Transaction databases vs. sequence databases
A transaction database
TID
10 20 30 40

A sequence database
SID
10 20 30 40

itemsets
a, b, d a, c, d a, d, e b, e, f

sequences
<a(abc)(ac)d(cf)> <(ad)c(bc)(ae)> <(ef)(ab)(df)cb> <eg(af)cbc>
3

Applications
Applications of sequential pattern mining
Customer shopping sequences:
First buy computer, then CD-ROM, and then digital camera, within 3 months.

Medical treatments, natural disasters (e.g., earthquakes), science & eng. processes, stocks and markets, etc. Telephone calling patterns, Weblog click streams DNA sequences and gene structures

Subsequence vs. super sequence

A sequence is an ordered list of events, denoted < e1 e2 el >
Given two sequences =< a1 a2 an > and =< b 1 b 2 bm > is called a subsequence of , denoted as , if there exist integers 1 j1 < j2 << jn m such that a1 bj1, a2 bj2,, an bjn is a super sequence of
E.g.=< (ab), d> and =< (abc), (de)>
5

What Is Sequential Pattern Mining?

Given a set of sequences and support threshold, find the complete set of frequent subsequences A sequence : < (ef) (ab) (df) c b >
A sequence database
SID 10 20 30 40 sequence <a(abc)(ac)d(cf)> <(ad)c(bc)(ae)> <(ef)(ab)(df)cb> <eg(af)cbc> An element may contain a set of items. Items within an element are unordered and we list them alphabetically.

<a(bc)dc> is a subsequence of <a(abc)(ac)d(cf)>

Given support threshold min_sup =2, <(ab)c> is a

sequential pattern

Challenges on Sequential Pattern Mining

A huge number of possible sequential patterns are hidden in databases
A mining algorithm should

find the complete set of patterns, when possible, satisfying the minimum support (frequency) threshold be highly efficient, scalable, involving only a small number of database scans
be able to incorporate various kinds of userspecific constraints
7

Studies on Sequential Pattern Mining

Concept introduction and an initial Apriori-like algorithm
Agrawal & Srikant. Mining sequential patterns, [ICDE95]

Apriori-based method: GSP (Generalized Sequential Patterns: Srikant

& Agrawal [EDBT96])

Pattern-growth methods: FreeSpan & PrefixSpan (Han et al.KDD00; Pei, et al. [ICDE01]) Vertical format-based mining: SPADE (Zaki [Machine Leanining00]) Constraint-based sequential pattern mining (SPIRIT: Garofalakis, Rastogi, Shim [VLDB99]; Pei, Han, Wang [CIKM02]) Mining closed sequential patterns: CloSpan (Yan, Han & Afshar

[SDM03])
8

Methods for sequential pattern mining

Apriori-based Approaches
GSP SPADE

Pattern-Growth-based Approaches
FreeSpan PrefixSpan

The Apriori Property of Sequential Patterns

A basic property: Apriori (Agrawal & Sirkant94)

If a sequence S is not frequent, then none of the super-sequences of S is frequent

E.g, <hb> is infrequent so do <hab> and <(ah)b>
Seq. ID 10 20 30 40 50 Sequence <(bd)cb(ac)> <(bf)(ce)b(fg)> <(ah)(bf)abf> <(be)(ce)d> <a(bd)bcb(ade)>

Given support threshold min_sup =2

10

GSPGeneralized Sequential Pattern Mining

GSP (Generalized Sequential Pattern) mining algorithm Outline of the method
Initially, every item in DB is a candidate of length-1 for each level (i.e., sequences of length-k) do
scan database to collect support count for each candidate sequence generate candidate length-(k+1) sequences from length-k frequent sequences using Apriori

repeat until no frequent sequence or no candidate can be found

Major strength: Candidate pruning by Apriori

11

Finding Length-1 Sequential Patterns

Initial candidates:
<a>, <b>, <c>, <d>, <e>, <f>, <g>, <h>

Scan database once, count support for candidates

min_sup =2
Seq. ID 10 20 30 40 50 Sequence <(bd)cb(ac)> <(bf)(ce)b(fg)> <(ah)(bf)abf> <(be)(ce)d> <a(bd)bcb(ade)>

Cand Sup <a> 3 <b> 5 <c> 4 <d> 3 <e> 3 <f> 2 <g> 1 <h> 1
12

Generating Length-2 Candidates

<a> <b> <ab> <bb> <cb> <db> <c> <ac> <bc> <cc> <dc> <d> <ad> <bd> <cd> <dd> <e> <ae> <be> <ce> <de> <f> <af> <bf> <cf> <df>

51 length-2 Candidates

<a> <b> <c> <d>

<aa> <ba> <ca> <da>

<e>
<f> <a> <a> <b> <(ab)> <c> <(ac)> <d> <(ad)>

<ea>
<fa> <e> <(ae)>

<eb>
<fb> <f> <(af)>

<ec>
<fc>

<ed>
<fd>

<ee>
<fe>

<ef>
<ff>

<b>
<c> <d> <e>

<(bc)>

<(bd)>
<(cd)>

<(be)>
<(ce)> <(de)>

<(bf)>
<(cf)> <(df)> <(ef)>

Without Apriori property, 8*8+8*7/2=92 candidates

<f>

Apriori prunes 13 44.57% candidates

Finding Lenth-2 Sequential Patterns

Scan database one more time, collect support count for each length-2 candidate There are 19 length-2 candidates which pass the minimum support threshold
They are length-2 sequential patterns

14

The GSP Mining Process

5th scan: 1 cand. 1 length-5 seq. pat. <(bd)cba>

Cand. cannot pass sup. threshold

Cand. not in DB at all 4th scan: 8 cand. 6 length-4 seq. <abba> <(bd)bc> pat. 3rd scan: 46 cand. 19 length-3 seq. <abb> <aab> <aba> <baa> <bab> pat. 20 cand. not in DB at all 2nd scan: 51 cand. 19 length-2 seq. <aa> <ab> <af> <ba> <bb> <ff> <(ab)> <(ef)> pat. 10 cand. not in DB at all 1st scan: 8 cand. 6 length-1 seq. <a> <b> <c> <d> <e> <f> <g> <h> pat. Seq. ID Sequence

min_sup =2

10

<(bd)cb(ac)>

20
30 40 50

<(bf)(ce)b(fg)>
<(ah)(bf)abf> <(be)(ce)d> <a(bd)bcb(ade)>
15

The GSP Algorithm

Take sequences in form of <x> as length-1 candidates Scan database once, find F1, the set of length-1 sequential patterns Let k=1; while Fk is not empty do
Form Ck+1, the set of length-(k+1) candidates from Fk; If Ck+1 is not empty, scan database once, find Fk+1, the set of length-(k+1) sequential patterns Let k=k+1;
16

The GSP Algorithm

Benefits from the Apriori pruning
Reduces search space

Bottlenecks
Scans the database multiple times

Generates a huge set of candidate sequences

There is a need for more efficient mining methods

17

The SPADE Algorithm

SPADE (Sequential PAttern Discovery using Equivalent Class) developed by Zaki 2001 A vertical format sequential pattern mining

method
A sequence database is mapped to a large set of Item: <SID, EID> Sequential pattern mining is performed by
growing the subsequences (patterns) one item at a

time by Apriori candidate generation

18

The SPADE Algorithm

19

Bottlenecks of Candidate Generate-and-test

A huge set of candidates generated.
Especially 2-item candidate sequence.

Multiple Scans of database in mining.

The length of each candidate grows by one at each
database scan.

Inefficient for mining long sequential patterns.

A long pattern grow up from short patterns An exponential number of short candidates
20

PrefixSpan (Prefix-Projected Sequential Pattern Growth)

PrefixSpan
Projection-based But only prefix-based projection: less projections and quickly shrinking sequences

J.Pei, J.Han, PrefixSpan : Mining sequential patterns efficiently by prefix-projected pattern growth. ICDE01.

21

Prefix and Suffix (Projection)

<a>, <aa>, <a(ab)> and <a(abc)> are prefixes

of sequence <a(abc)(ac)d(cf)>
Given sequence <a(abc)(ac)d(cf)>
Prefix Suffix (Prefix-Based Projection)

<a> <aa> <ab>

<(abc)(ac)d(cf)> <(_bc)(ac)d(cf)> <(_c)(ac)d(cf)>

22

Mining Sequential Patterns by Prefix Projections

Step 1: find length-1 sequential patterns
<a>, <b>, <c>, <d>, <e>, <f>

Step 2: divide search space. The complete set of seq. pat. can be partitioned into 6 subsets:
The ones having prefix <a>; The ones having prefix <b>; The ones having prefix <f>
SID 10 sequence <a(abc)(ac)d(cf)>

20
30 40

<(ad)c(bc)(ae)>
<(ef)(ab)(df)cb> <eg(af)cbc>
23

Finding Seq. Patterns with Prefix <a>

Only need to consider projections w.r.t. <a>
<a>-projected database: <(abc)(ac)d(cf)>, <(_d)c(bc)(ae)>, <(_b)(df)cb>, <(_f)cbc>

Find all the length-2 seq. pat. Having prefix <a>: <aa>, <ab>, <(ab)>, <ac>, <ad>, <af>
Further partition into 6 subsets
Having prefix <aa>;
Having prefix <af>
SID sequence

10
20 30 40

<a(abc)(ac)d(cf)>
<(ad)c(bc)(ae)> <(ef)(ab)(df)cb> <eg(af)cbc>
24

Completeness of PrefixSpan
SDB
SID sequence

10
20 30 40

<a(abc)(ac)d(cf)>
<(ad)c(bc)(ae)> <(ef)(ab)(df)cb> <eg(af)cbc>

Length-1 sequential patterns <a>, <b>, <c>, <d>, <e>, <f>

Having prefix <a> <a>-projected database <(abc)(ac)d(cf)> <(_d)c(bc)(ae)> <(_b)(df)cb> <(_f)cbc>

Having prefix <c>, , <f> <b>-projected database

Having prefix <b>

Length-2 sequential patterns <aa>, <ab>, <(ab)>, <ac>, <ad>, <af>

Having prefix <aa> Having prefix <af>

<aa>-proj. db

<af>-proj. db
25

The Algorithm of PrefixSpan

Input: A sequence database S, and the minimum support threshold min_sup Output: The complete set of sequential patterns Method: Call PrefixSpan(<>,0,S) Subroutine PrefixSpan(, l, S|) Parameters:
: sequential pattern, l: the length of ; S|: the -projected database, if <>; otherwise; the sequence database S
26

The Algorithm of PrefixSpan(2)

Method
1. Scan S| once, find the set of frequent items b such that:
a) b can be assembled to the last element of to form a sequential pattern; or b) <b> can be appended to to form a sequential pattern.

2. For each frequent item b, append it to to form a sequential pattern , and output ; 3. For each , construct -projected database S|, and call PrefixSpan(, l+1, S|).
27

Efficiency of PrefixSpan
No candidate sequence needs to be generated Projected databases keep shrinking Major cost of PrefixSpan: constructing

projected databases
Can be improved by bi-level projections
28

Optimization in PrefixSpan
Single level vs. bi-level projection
Bi-level projection with 3-way checking may reduce the number and size of projected databases

Physical projection vs. pseudo-projection

Pseudo-projection may reduce the effort of projection when the projected database fits in main memory

Parallel projection vs. partition projection

Partition projection may avoid the blowup of disk space

29

Scaling Up by Bi-Level Projection

Partition search space based on length-2 sequential patterns Only form projected databases and pursue recursive mining over bi-level projected databases

30

Speed-up by Pseudo-projection
Major cost of PrefixSpan: projection
Postfixes of sequences often appear
repeatedly in recursive projected databases

When (projected) database can be held in main memory, use pointers to form projections
Pointer to the sequence Offset of the postfix s=<a(abc)(ac)d(cf)> <a> s|<a>: ( , 2) <(abc)(ac)d(cf)> <ab> s|<ab>: ( , 4) <(_c)(ac)d(cf)> 31

Pseudo-Projection vs. Physical Projection

Pseudo-projection avoids physically copying postfixes
Efficient in running time and space when database can be held in main memory

However, it is not efficient when database cannot fit in main memory

Disk-based random accessing is very costly

Suggested Approach:
Integration of physical and pseudo-projection Swapping to pseudo-projection when the data set fits in memory
32

Performance on Data Set C10T8S8I8

33

Performance on Data Set Gazelle

34

Effect of Pseudo-Projection

35

CloSpan: Mining Closed Sequential Patterns

A closed sequential pattern s: there exists no superpattern s such that s s, and s and s have the same support

Motivation: reduces the number of (redundant) patterns but attains the same expressive power
Using Backward Subpattern and Backward Superpattern pruning to prune redundant search space

36

CloSpan: Performance Comparison with PrefixSpan

37

Constraints for Seq.-Pattern Mining

Item constraint
Find web log patterns only about online-bookstores

Length constraint
Find patterns having at least 20 items

Super pattern constraint

Find super patterns of PC digital camera

Aggregate constraint
Find patterns that the average price of items is over $100

38

More Constraints
Regular expression constraint
Find patterns starting from Yahoo homepage, search for hotels in Washington DC area Yahootravel(WashingtonDC|DC)(hotel|motel|lodging)

Duration constraint
Find patterns about 24 hours of a shooting

Gap constraint
Find purchasing patterns such that the gap between each consecutive purchases is less than 1 month
39

From Sequential Patterns to Structured Patterns

Sets, sequences, trees, graphs, and other structures
Transaction DB: Sets of items
{{i1, i2, , im}, }

Seq. DB: Sequences of sets:

{<{i1, i2}, , {im, in, ik}>, }

Sets of Sequences:
{{<i1, i2>, , <im, in, ik>}, }

Sets of trees: {t1, t2, , tn} Sets of graphs (mining for frequent subgraphs):
{g1, g2, , gn}

Mining structured patterns in XML documents,

40

Episodes and Episode Pattern Mining

Other methods for specifying the kinds of patterns
Serial episodes: A B Parallel episodes: A & B

Regular expressions: (A | B)C*(D E)

Methods for episode pattern mining

Variations of Apriori-like algorithms, e.g., GSP

Database projection-based pattern growth

Similar to the frequent pattern growth without candidate generation
41

Periodicity Analysis
Periodicity is everywhere: tides, seasons, daily power consumption, etc. Full periodicity
Every point in time contributes (precisely or approximately) to the periodicity

Partial periodicit: A more general notion

Only some segments contribute to the periodicity
Jim reads NY Times 7:00-7:30 am every week day

Cyclic association rules

Associations which form cycles

Methods
Full periodicity: FFT, other statistical analysis methods Partial and cyclic periodicity: Variations of Apriori-like mining methods

42

Summary
Sequential Pattern Mining is useful in many application, e.g. weblog analysis, financial market prediction, BioInformatics, etc. It is similar to the frequent itemsets mining, but with consideration of ordering. We have looked at different approaches that are descendants from two popular algorithms in mining frequent itemsets
Candidates Generation: AprioriAll and GSP Pattern Growth: FreeSpan and PrefixSpan
43