J. Intelligent Learning Systems & Applications, 2010, 2, 117-125
doi:10.4236/jilsa.2010.23015 Published Online August 2010 (http://www.SciRP.org/journal/jilsa)
117
A SOM-Based Document Clustering Using
Frequent Max Substrings for Non-Segmented
Texts
Todsanai Chumwatana, Kok Wai Wong, Hong Xie
School of Information Technology, Murdoch University, South St, Murdoch, Australia.
Received March 25th, 2010; revised July 15th, 2010; accepted July 30th, 2010.
ABSTRACT
This paper proposes a non-segmented document clustering method using self-organizing map (SOM) and frequent max
substring technique to improve th e efficiency of information retrieval. SOM has been wid ely used for document cluster-
ing and is successful in many applications. However, when applying to non-segmented document, the challenge is to
identify any interesting pattern efficien tly. There are two main phases in the propose method: p reprocessing phase and
clustering phase. In the preprocessing phase, the frequent max substring technique is first applied to discover the pat-
terns of interest called Frequent Max substrings that are long and frequent substrings, rather than individual words
from the non-segmented texts. These discovered patterns are then used as ind exing terms. The indexing terms together
with their number of occurrences form a document vector. In the clustering phase, SOM is used to generate the docu-
ment cluster map by using the feature vector of Frequent Max substrings. To demonstrate the proposed technique, ex-
perimental studies and comparison results on clustering the Thai text documents, which consist of non-segmented texts,
are presented in this paper. The results show that the proposed technique can be used for Thai texts. The document
cluster map generated with the method can be used to find the relevant documents more efficiently.
Keywords: Frequent Max Substring, Self-Organizing Map, Document Clustering
1. Introduction
Document clustering has been an important issue [1] due
to the rapid growth in the number of electronic docu-
ments. Document clustering, sometimes can be general-
ized as text clustering, indentifies the similarity of docu-
ments and summarize a large number of documents using
key attributes of the clusters. Document clustering uses
unsupervised learning techniques and may assist fast
information retrieval or filtering [2]. This is because clus-
tering technique categorizes documents into groups
based on their similarity in term of their member occur-
rences. Thus clustering can be used to categorize docu-
ment databases and digital libraries, as well as providing
useful summary information of the categories for brows-
ing purposes. In information retrieval, a typical search on
document database or the World Wide Web can return
several thousands of documents in response to the user’s
queries. It is often very difficult for users to identify their
documents of interest from such a huge number of
documents. Clustering the documents enables the user to
have a clear and easy grasp of the relevant documents
from the collection of documents which are similar to
each other and could be rel evant to the user’s queries.
For text clustering in information retrieval, a document
is normally considered as a bag of words, even though a
document actually consists of a sequence of sentences
and each sentence is composed from a sequence of words.
Very often the positions of words are ignored when per-
forming document clustering. Words, also known as in-
dexing terms, and their weights in documents are usually
used as important parameters to compute the similarity of
documents [3]. Those documents that contain similar
indexing terms and frequencies will be grouped under th e
same cluster. This process is straightforward for Euro-
pean languages where words are clearly defined by word
delimiter such as space or other special symbols. Euro-
pean texts are explicitly segmented in to word tokens that
are used as indexing terms. Many algorithms have been
developed to calculate the similarity of documents and to
build clusters for fast information retrieval [4]. In con-
trast, document clustering can be a challenging task for

A SOM-Based Document Clustering Using Frequent Max Substrings for Non-Segmented Texts 118
many Asian languages such as Chinese, Japanese, Ko-
rean and Thai, because these languages are non-segmented
languages, i.e., a sentence is written continuously as a
sequence of characters without explicit word boundary
delimiters. Due to this characteristic, texts in a non-seg-
mented document cannot be directly used to calculate the
similarity. Some preprocessing needs to be performed
first to discover keywords for Asian documents before
clustering. As a result, most approaches for clustering
non-segmented documents consist of two phases: a text
mining process to extract the keywords, and a document
clustering process to compute the similarity between the
input documents.
2. Keyword Extraction
Keywords are usually regarded as an important key to
identifying the main content of the documents. Most of
the semantics are usually carried by nouns, although a
sentence in a natural language text is composed of nouns,
pronouns, articles, verbs, adjectives, adverbs, and con-
nectives. Keyword extraction is one of the main applica-
tions of text mining. The objective of text mining is to
exploit useful information or knowledge contained in
textual documents [5]. Information Extraction (IE) is an
essential task in text mining that describes a process of
discovering interesting keywords underlying unstruc-
tured natural-language texts. Most keyword extraction
methods proposed in the literature were accomplished by
constructing a set of words from given texts. Keywords
will then be selected from the set of words during the
preprocessing step. Many approaches have been pro-
posed to extract keywords from non-segmented docu-
ments such as Chinese [6], Japanese [7] or Thai docu-
ments [8]. Most techniques are based on word segmenta-
tion which is one of the most widely used information
extraction techniques in Natural language Processing
(NLP). However, most word segmentation approaches
involve complex language analysis and require long
computational time. After keyword extraction is per-
formed, keywords are then transformed into feature vec-
tor of the words that appear in the documents. The
term-weights (usually term-frequencies) of the words are
also contained in each feature vector. The vector space
model (VSM) has been a standard model of representing
documents by containing the set of words with their fre-
quencies [1]. In the VSM, each document is replaced by
the vector of the words. The vector size is dependent on
the number of keywords that appear in the documents.
For instance, let wik be the weight of keyword k that ap-
pear in the document i, and Di = (wi1, wi2,…, wit) is the
feature vector for document i, where t is the number of
unique words of all documents. Therefore, the size of
the feature vector is equal to t dimension as shown in
Figure 1.
Figure 1. The example of the document vectors in 3-dimen-
sion
From Figure 1, the similarity between two documents
can be computed with one of several similarity measures
based on two corresponding feature vectors, e.g., cosine
measure, Jaccard measure, and Euclidean distance meas-
ure [9].
3. Document Clustering Algorithms
In document clustering, there are two main approaches:
hierarchical and partitional approaches [10,11,4]. The
hierarchical approach produces document clusters by
using a nested sequence of partitions that can be repre-
sented in the form of a tree structure called a dendrogram.
The root of the tree contains one cluster covering all data
points, and singleton cluster of individual data point are
shown on the leaves of the tree. There are two basic
methods when performing hierarchical clustering: ag-
glomerative (bottom up) and divisive (top down) clus-
tering [4]. The advantages of hierarchical approach are
that it can take any form of similarity function, and also
the hierarchy of clusters allows users to discover clusters
at any level of detail. However, this technique may suffer
from the chain effect, and its space requirement is at least
quadratic or O(n2) compared to the k-means algorithm
that provide O(Iknm) where I is the number of necessary
iterations, k is the number of clusters, n is the number of
documents and m is the dimensionality of the vectors.
The partitional approach [12], on the other hand, can be
divided into several techniques, e.g., k-means [13], Fuzzy
c-means [14], QT (quality threshold) [15] algorithms., Th e
k-means algorithm is more widely used among all clus-
tering algorithms because of its efficiency and simplicity.
The basic idea of k-means algorithm is that it separates a
given data into k clusters where each cluster has the cen-
ter point, also called centroid, which can be used to rep-
resent the cluster. The main advantages of k-means algo-
rithm are its efficiency and simplicity. Its weaknesses are
that it is only applicable to data sets where the notion of
the mean is defined, the number of clusters can be iden ti-
fied by users, and it is sensitive to data points that are
very far away from other points called outliers [1]. Fur-

A SOM-Based Document Clustering Using Frequent Max Substrings for Non-Segmented Texts 119
thermore, Self-organizing map (SOM) [16,17] can be
used as one of the clustering algorithms in the family of
an artificial neural network. The self-organizing map is
unsupervised neural network, capable of ordering high
dimensional data in such a way that similar inputs are
grouped spatially close to each other. To use SOM in
document clustering, text documents are described by
features with high dimensionality, and SOM based tech-
niques have been successfully applied to document clus-
tering. Some of the successful applications of SOM in
document clustering are described in the next section.
4. Related Works
Many clustering techniques have been developed and can
be applied to clustering English documents Most of these
traditional approaches use documents as the basis for
clustering [18,19]. The Vector Space Document (VSD)
model is a very widely used data representation model
for document clustering [20]. This data model starts with
a representation of any document as a feature vector of
the words that appear in documents. The term-weights of
the words are also contained in the feature vectors. The
similarity measures are used to co mpute the similarity of
two document vectors. An alternative approach of docu-
ment clustering is phrase-based document clustering.
Zamir and Etzioni [21] introduced the notion of phrase-
based document clustering. They proposed to use a gen-
eralized suffix-tree to obtain information about the phrases
between two documents and use common phrases to
cluster the documents. According to [22], Bakus, Hussin,
and Kamel used a hierarchical phrase grammar extraction
procedure to identify phrases from documents and used
these phrases as features for document clustering. The
self-organizing map (SOM) method was used as the
clustering algorithm. An improvement in clustering per-
formance was demonstrated when using phrases rather
than single words as features.
Mladenic and Grobelnik used a Naive Bayesian method
to classify documents based on word sequences of dif-
ferent length [23]. Experimental results show that using
the word sequences whose length is no more than 3
words can improve the performance of a text classifica-
tion system. But when the average length of used word
sequences is longer than 3 words, there will be no dif-
ference between usin g w or d sequences or single words.
However, there are not many research works on phrase-
based document clustering for Asian languages, primar-
ily due to the fact that most Asian langu age texts are non-
segmented and it is difficult to separate words and
phrases from the non-segmented texts. Most document
clustering approaches require a preprocessing stage
where word segmentation, stopwo rd removal or semantic
analysis are performed. NLP techniques provide good
support for this step. Word segmentation is very impor-
tant step involved in most NLP processing tasks. A text
is separated into a sequence of tokens by using word
segmentation techniques. Many approaches have been
proposed for Asian languages such as Chinese, Japanese,
Korea and Thai languages.
In [24], a Chinese document clustering method using
data mining technique and neural network model was
proposed. This technique was divided into two main
parts: the preprocessing part which provides Chinese
sentence segmentation method, and the clustering part
that adopts the dynamical SOM model with a view to
dynamically clustering data. In addition, this method uses
term vectors clustering process instead of document vec-
tors clustering process.
In Thai language, Canasai and Chuleerat propose a
parallel algorithm for clustering text documents based on
spherical k-means [25]. They implemented an algorithm
on the PIRUN Linux Cluster, which is a parallel com-
puter using cluster computing technology. Experimental
results show that the use of parallel algorithm can sig-
nificantly improve clustering performance.
5. A SOM Based Clustering Using Frequent
Max Substrings for Non-Segmented Texts
In this section, we describe a new method that combines
Kohonen’s SOM and frequent max substring technique to
process the non-segmented text documents into clusters.
SOM is one of the main unsupervised learning methods
in the family of artificial neural networks (ANN) that
was first developed by Teuvo Kohonen in 1984 [26]. The
SOM can be visualized as a regular two-dimensional
array of cells or nodes (neurons). The SOM algorithm
defines a mapping from the input vector onto a two-
dimensional array of nodes. When the input vector x(t)

Rn is given, it is connected to all neurons in the SOM
array denoted as vector mi(t) Rn ,whic h are associated
by each neuron and is gradually modified in the learning
process. The input vector x(t) Rn is the inp ut data sets
where t is the indexing terms of the input documents.
These input data sets have to be mapped with all neurons
in the map that is denoted as two-dimensional network of
cells or the model vector mi(t) Rn.



In mapping, the node where vector mi is most similar
to the input vector x will be activ ated. This node is often
called a best-matching node or a winner. The winner and
a number of its neighboring nodes in the SOM array are
then turned towards the input vector x according to the
learning principle.
The frequent max substring technique is an information
extraction technique used to identify the terms called
frequent max substrings from non-segmented texts where
the word boundary and characteristic are not clearly de-
fined. This technique was first introduced in 2008 [27]
and has been proposed as an alternative language-indepen-
dent technique for keyword extraction for non-segmented
texts [28,29]. It also has been applied in application for

A SOM-Based Document Clustering Using Frequent Max Substrings for Non-Segmented Texts 120
indexing for non-segmented languages [8,30] as this
technique provides good significant in term of the effi-
ciency of storage space which could be able to support
the rapid growth in the number of electronic non-segmen-
ted documents. The frequent max substring technique
classifies indexing terms, known as frequent max sub-
strings, from the non-segmented texts where the word
boundaries are not clearly defined. The frequent max
substrings refer to the substrings that appear frequently
(at a given frequency threshold value f) and have the
maximum length on the given string, so these terms are
likely to be the patterns of interest. We extract the set of
frequent max substrings or FMAX by using the frequent
max substring technique. This technique uses Frequent
Suffix Trie or FST data structure to explore the indexing
terms [27]. The FST data structure is similar to suffix trie
structure, that is an efficient substring enumeration method.
However, FST data structure enumerates substrings with
their frequencies and positions information while suffix
trie structure enumerates only substrings without their
frequencies information. Therefore, we employ FST data
structure in order to support extracting the frequent max
substrings. In this technique, the parameter and the pre-
determined frequency are applied to reduce the number
of the indexing terms. This method uses the two reduc-
tion rules: 1) reduction rule using the predetermined fre-
quency to check extracting termination, 2) reduction rule
using superstring definition to reduce the number of in-
dexing terms extracted. This technique also uses heap
data structure to support computation [27].
In this paper, we use a set of non-segmented docu-
ments (Thai documents) as an input to train a map using
SOM. We will describe the process of clustering as fol-
lows.
Let D be a document collection consisting of n docu-
ments, d1, d2, ..., dn. Firstly, we use the frequent max
substring technique [27] to generate a set of frequent max
substrings at the given frequency threshold value f from
the document collectio n to be used as the set of indexing
terms for the document collection.
Assuming the above process produces m frequent max
substrings from the document collection, denoted FMAX
= (fm1, fm2, ..., fmm). where fmi is the ith frequent max
substring generated from the document collection. These
m substrings are used as our indexing terms.
We will then calculate th e weight wij which represents
the frequency of indexing term fmi occurring in docu-
ment dj for each indexing term and each document. Fi-
nally we construct an m × n matrix of such weights. In
this matrix, row i represents the frequencies of occur-
rence of the ith indexing term fmi in the n documents,
while jth column represents the document vector for
document j, as depicted in Figure 2.
Figure 2 shows an example of document matrix. In a
document matrix, each element wij is at least at f if fmi
occurs in the document dj or 0 if fmi does not appear in
the document dj, i.e.,
if occu r s in
0 otherwise
ij
ij
f
fm d
w





After the document matrix is obtained, the document
vectors are presented to SOM for clustering. These
documents can be labeled into neurons according to the
similarity of their document vectors. Two documents
containing the same or similar document vectors will
map to the same neuron. In contrast, the documents may
map to distant neurons if they contain different or non-
overlapping frequent max substrings. Furthermore, the
documents with similar frequent max substrings may
map to neighbouring neurons. This means that the neu-
rons can form the document clusters by examining
mapped neurons in the document cluster map. We depict
the organization of the document map that clusters simi-
lar documents into the same neuron as shown in the
boxes. fms in the boxes represent the content of docu-
ments in the collection .
Figure 2. The example of the document matrix at the given
frequency threshold value f is equal to 2
Figure 3. The example of the document cluster map

A SOM-Based Document Clustering Using Frequent Max Substrings for Non-Segmented Texts 121
After the SOM has been trained , the docu men t clus ters
are formed by labeling each neuron that contains certain
documents of similar type. The documents in the same
neuron may not contain exactly the same set of frequent
max substrings or FMAX, but they usually contain mostly
overlapping frequent max substrings. As a result, the
document cluster map can be used as a prediction model
to generate the different groups of similar documents,
and each group will then be used to specify the document
type by comparing frequent max substrings of each
group wi th keywords of each area. In Figure 4, we depi ct
clustering the documents into different groups, by map-
ping an input data with neurons in the document cluster
map to find the document groups of several types.
From Figure 4, the following will describe the process
of matching an input data x with the neurons in the
document cluster map by using SOM.
Let us consider the input vector x = [x1, x2, …, xn]t

Rn as the input data sets where t is the frequent max sub-
strings of the input documents. These input data sets
have to be matched with all neurons in the map that is
denoted as two-dimensional network of cells or the
model vector mi = [mi1, mi2, …, min]t R
n depicted in
Figure 5. Each neuron i in the network contains the
model vector mi, which has the same number of indexing
terms as the input vector x.

From Figure 5, the input vector x is compared with all
neurons in the model vector mi to find the best matching
node called the winner. The winner unit is the neuron on
the map where the set of the frequent max substrings of
the input vector x is the same or similar to the set of the
frequent max substrings of the model vector mi by using
some matching criterion e.g. the Euclidean distances
between x and mi. As a result, this method can be used to
cluster documents into different groups, and also sug-
gested that this can use to reduce the search time for the
relevant document.
6. Experimental Studies and Comparison
Results
In this section, we describe an experiment for clustering
non-segmented documents (Thai documents) based on
the proposed SOM and frequent max substring technique.
We also compare the proposed technique with the SOM
based documents clustering using single words as fea-
tures and hierarchical clustering technique using single
words on a group of documents. 50 Thai documents were
used as an input dataset to train a map. All Thai docu-
ments used are from Thai news websites that consist of
different categories of contents: sport, travel, education
and political news. The documents h ave varying lengths.
The set of documents contains 103,287 characters, and
average document length is 2,065 characters or 78 words
per document. The basic statistics for the text collection
are shown in Table 1.
Figure 4. Neuron network architecture
Figure 5. Self-organizing map
Table 1. Basic statistics for Thai text collection
No. of
Docs No. of Chars No. of Words
Avg.
Chars./
Docs
Avg.
Words/
Docs
Sport news15 24727 997 1648.4666.46
Travel news15 29096 1022 2078.2873
Political
news 15 38017 1398 2534.4693.2
Education
news 5 9445 336 1889 67.2
In our proposed technique, the set of frequent max
substring was first generated by frequent max substring
technique at the given frequency threshold value, which
is equal to 2 from the document dataset and 35 frequent
max substrings, the long and frequently occurring terms
in sport, travel, political and education documents, were
used as the set of indexing terms for this document col-
lection. The 50 input documents are then transformed to
a document matrix of weighted frequent max substring
occurrence. Hence, these 35 indexing terms and 50 input
documents to form a 50 × 35 matrix, where each docu-
ment vector was represented by each column of the ma-
trix, and the rows of the matrix correspond to the index-
ing terms. We use this 50 × 35 matrix to train a map us-
ing SOM, and the number of neurons was set as 9 in
SOM program as shown in Figure 6.
In the experimental study, 9 was set as the number of
neurons because the several numbers of neurons were
investigated, and 9 neurons provided the best result among
them. From experimental studies, the group of political,
sport, and education documents provided good results as

A SOM-Based Document Clustering Using Frequent Max Substrings for Non-Segmented Texts 122
similar documents of each type were mapped onto the
neuron. It can be observed that some errors occurred
within the group of travel documents. The travel docu-
ments were mapped onto several neurons due to overlap-
ping terms that appeared across different type of docu-
ments.
The Figure 6 showed the map containing 9 neurons
and 50 Thai documents. Each neuron contains the group
of similar documents.
From Figure 6, the experimental result showed that
SOM can cluster 50 documents into 5 neurons on the
map, and the similar documents were grouped into the
same neuron as shown in Table 2.
As observed from the results, this technique can be
used to cluster non-segmented documents into several
groups according to their similarity. The accuracy of this
technique is up to 83.25%. However, from this experi-
ment, we have found that the groups of education and
sport documents are mapped onto the same neuron
(Neuron5) because they both contain mostly overlapping
frequent max substrings such as ผลการแขงขัน (competi-
tion result), การจัดอันดับ (position ranking), ไดรับรางวัล
(getting award), etc. Furthermore, the contents of docu-
ments and generated indexing terms are also the main
factors that impact the accurate value. The content of one
document may have overlapping terms from two differ-
ent types of documents. For instance, Education5, Travel
1, Travel 3, Travel 11 and Travel 14 documents are
mapped onto the neuron 3 because they are presenting
information on ecotourism, containing overlapping terms
from education and travel documents.
The methodology to measure the clustering ap-
proaches is to compare the group of documents occur-
rences. In this paper, we compare our technique with the
SOM based documents clustering using single words [31]
as features and hierarchical based document clustering
using single words [4,12]. In the SOM based documents
clustering using single words, single words were first
extracted by using Thai word segmentation from the
same document data set and 35 single words, most fre-
quent occurring single keywords in education, sport,
travel and political documents, and 50 input documents
are used to form a 50 × 35 matrix, which is the same ma-
trix size as our earlier experiment to train a map using
SOM. From experimental studies, it can be observed that
only the groups of politic, and travel documen ts provided
fairly good results as shown in Table 3.
The group of similar travel documents was mapped
onto the neuron as well as the group of political docu-
ments. Meanwhile, the education and sport documents
were distributed onto several neurons because most sin-
gle words extracted from education and sport documents
are general terms. These general terms can be shared in
many types of documents.
To depict the experiment results, Figure 7 showed the
Figure 6. SOM contains 9 neurons and the group of similar
documents from 50 Thai document collection
Figure 7. SOM contains 9 neurons and the group of similar
documents from 50 Thai document collection by using sin-
gle words as features
map containing 9 neurons and 50 Thai documents, and
the documents were grouped into the neurons as shown
in Table 3.
It can be observed that the accuracy of the SOM based
documents clustering using single words is 72.21%.
Moreover, from the experiment results, we have found
that the SOM based documents clustering using frequent
max substrings provides better result than the SOM
based documents clustering using single words because
the frequent max substrings can be used to describe the

A SOM-Based Document Clustering Using Frequent Max Substrings for Non-Segmented Texts 123
Table 2. Clustering results of using SOM and frequent max substring technique
Neuron ID Row Column Document ID
Neuron 5 1 2
Political 1, Education 1, Education 2, Education 3, Education
4, Sport 1, Sport 2, Sport 3, Sport 4, Sport 5, Sport 6, Sport 7,
Sport 8, Sport 9, Sport 10, Sport 11, Sport 12, Sport 13, Sport
14, Sport 15, Travel 10
Neuron 2 2 1 Political 2, Political 3, Political 4, Political 5, Political 6, Political 7, Political 8, Political 9, Political 10, Political 11,
Political 12, Political 13, Political 14, Political 15,
Neuron 4 2 2 Travel 2, Travel 4, Travel 5, Travel 6, Travel 7, Travel 8,
Travel 9, Travel 13, Travel 15
Neuron 1 3 1 Travel 12
Neuron 3 3 2 Education 5, Travel 1, Travel 3, Travel 11, Travel 14
Table 3. Clustering results of SOM based documents clustering using single words
Neuron ID Row Column Document ID
Neuron 2 1 1 Education 1, Sport 1, Sport 2, Sport 5, Sport 6, Political 1, Political 2, Political 3, Political 4, Political 6, Political 8,
Political 9, Political 10, Political 11, Political 12
Neuron 5 1 2 Education 3, Education 5, Sport 7, Sport 8, Sport 9, Sport 11,
Sport 13, Travel 7
Neuron 6 1 3 Sport 12, Sport 14,
Neuron 1 2 1 Sport 15, Travel 15, Politica l7
Neuron 4 2 2
Education 2, Sport 3, Sport 4, Travel 1, Travel 2, Travel 3, Travel 4,
Travel 5, Travel 6, Travel 8, Travel 9, Travel 10, Travel 11,
Travel 12, Travel 13, Travel 14, Political 13, Political 14,
Political 15
Neuron 3 3 2 Education 4, Sport 10, Politcal 5
content of the documents more specifically than using
single words, as the frequent max substrings can be re-
ferred to the frequently and long terms rather than indi-
vidual words. Moreover, many researches have also shown
an improvement in clustering performance when using
phrases rather than single words as features [21,22].
Additionally, we also compare our technique with the
hierarchical based document clustering using single
words. The hierarchical based document clustering is
used because it has been widely used and has been ap-
plied successfully in many applications in the area of
document clustering [12]. This method has also been
used to perform Thai documents clustering. To use this
technique with Thai language, single words were first
extracted by using Thai word segmentation techniques
from the same document dataset used in our experiment
that is discussed earlier. After word segmentation is per-
formed, single words are then transformed into feature
vector of the words that appear in documents. The
term-frequencies of the words are also contained in each
feature vector. The feature vector of the words is then
used to compute the similarity of the documents by using
the hierarchical clustering approach. In the hierarchical
clustering program, the feature vector of the words with
their frequencies was used as an input data, and the
number of clusters was set to 9 after trial-and-error. From
the experimental results, the group of political, travel and
education documents provided good results. However,
travel and education documents were grouped into the
same cluster (cluster 1). It can be observed that the travel
and education documents are sharing overlapping words
that appeared across different type of documents. In ad-
dition, some of travel, education, sport and travel docu-
ment were distributed across several small clusters. In
Ta bl e 4, the experimental result showed that the hierar-
chical clustering program can cluster 50 documents into
9 clusters.
As can be observed from the results, the accuracy of
the proposed method is up to 83.25%, meanwhile using
the SOM based documents clustering using single words
and the hierarchical clustering approach provide the ac-
curacies 72.21% and 79.75% respectively. The hierarchi-

A SOM-Based Document Clustering Using Frequent Max Substrings for Non-Segmented Texts 124
Table 4. Clustering results of using the hierarchical clus-
tering approach
Cluster ID Document ID
Cluster 1
Education 1, Education 2, Education 3, E du-
cation 4, Sport 9, Sport 13, Travel 1, Travel
2, Travel 3, Travel 4, Travel 5, Travel 7,
Travel 8, Travel 9, Travel 10, Travel 11,
Travel 12, Travel 13, Travel 14, Travel 15,
Political 15
Cluster 2 Education 5
Cluster 3 Sport 1
Cluster 4 Sport 2, Sport 3, Sport 4, Sport 5, Sport 6,
Sport 7, Sport 8, Sport 15
Cluster 5 Sport 10, Sport 11
Cluster 6 Sport 12
Cluster 7 Sport 14
Cluster 8 Travel 6
Cluster 9
Political 1, Political 2, Political 3, Political 4,
Political 5, Political 6, Political 7, Political 8,
Political 9, Political 10, Political 11, Political
12, Political 13, Political 14
cal clustering approach also created many small clusters
that containing only a few documents. As a result, an
improvement was demonstrated using frequent max sub-
strings rather than single words as features. This pro-
posed technique also does not require any pre-processing
technique to extract the frequent max substrings. Mean-
while, the SOM based documents clustering using single
words and the hierarchical based document clustering
using single words require word segmentation to extract
the single words.
7. Conclusions
This paper describes a non-segmented document cluster-
ing method using self- organizing map (SOM) and fre-
quent max substring technique to improve the efficiency
of information retrieval. We first use the frequent max
substring technique to discover pattern s of interest, called
frequent max substrings, rather than individual words
from Thai text documents, and these frequent max sub-
strings are then used as indexing terms with their number
of occurrences to form a document vector. SOM is then
applied to generate the document cluster map by using
the document vector. The experiment studies and com-
parison results on clustering the 50 Thai text documents
is presented in this paper. We compare the proposed
technique with the SOM based documents clustering
using single words and hierarchical based document
clustering technique with the use of single words for
grouping the document occurrences. From the experi-
mental results, our technique can be used to cluster 50
Thai documents into different clusters with more accu-
racy than using the SOM based documents clustering
using single words and the hierarchical clustering ap-
proaches. As a result, the generated document cluster
map from our technique can be used to find the relevant
documents according to a user’s query more efficiency.
REFERENCES
[1] B. Liu, “Web Data Mining: Exploring Hyperlinks,
Con-tents, and Usage Data,” 1st Edition, Springer-Verlag,
New York Berlin Heidelberg, 2007.
[2] D. R. K. R. D. Cutting, J. O. Pedersen, J. W. Tukey,
“Scatter/Gather: A Cluster-based Approach to Browsing
Large Document Collections,” Proceedings of ACM Spe-
cial Interest Group on Information Retrieval ‘92, Copen-
hague, 1992, pp. 318-329.
[3] I. Matveeva, “Document Representation and Multilevel
Measures of Document Similarity,” Irina Matveeva,
Document representation and multilevel measures of
document similarity, Proceedings of the 2006 Conference
of the North American Chapter of the Association for
Computational Linguistics on Human Language Tech-
nology: companion volume: doctoral consortium, New
York, 2006, pp. 235-238.
[4] G. K. M. Steinbach and V. Kumar, “A Comparison of
Docu-ment Clustering Techniques,” KDD Workshop on
Text Mining, Boston, 2000.
[5] A.-H. Tan, “Text Mining: The state of the art and the
challenges,” Proceedings of the PAKDD Workshop on
Knowledge Discovery from Advanced Databases, Beijing,
1999, pp. 65-70.
[6] Q. L. H. Jiao and H.-B. Jia, “Chinese Key word Extraction
Based on N-Gram and Word Co-occurrence, 2007 Inter-
national Conference on Computational Intelligence and
Security Workshops (CISW 2007), Harbin, 2007, pp.
124-127.
[7] J. Mathieu, “Adaptation of a Keyphrase Extractor for
Japanese Text,” Proceedings of the 27th Annual Confer-
ence of the Canadian Association for Information Science
(CAIS-99), Sherbrooke, Quebec, 1999, pp. 182-189.
[8] T. Chumwatana, K. W. Wong and H. Xie “An Automatic
Indexing Technique for Thai Texts Using Frequent Max
Substring,” 2009 Eight International Symposium on
Natural Language Processing, Bangkok, 2009, pp. 67-72.
[9] R. Feldman and J. Sanger, “The Text Mining Handbook:
Advanced Approaches in Analyzing Unstructured Data,”
Cambridge University Press, Cambridge, 2006.
[10] A. K. Jain and R. C. Dubes, “Algorithms for Clustering
Data,” Prentice Hall, New Jersey, 1988.
[11] L. Kaufman and P. J. Rousseeuw, “Finding Groups in
Data: An Introduction to Cluster Analysis,” John Wiley
and Sons, New York, 1990.
[12] G. K. Y. Zhao, “Comparison of Agglomerative and Parti-
tional Document Clustering Algorithms,” The SIAM
workshop on Clustering High-dimensional Data and Its
Applications, Washington, DC, April 2002.

A SOM-Based Document Clustering Using Frequent Max Substrings for Non-Segmented Texts 125
[13] Z. Huang, “Extensions to the K-means Algorithm for
Clustering Large Datasets with Categorical Values,” Data
Mining and Knowledge Discovery, Vol. 2, No. 3, 1998,
pp. 283-304.
[14] D. Dembele and P. Kastner, “Fuzzy C-Means Method for
Clustering Microarray Data,” Bioinformatics, Vol. 19, No.
8, 2003, pp. 973-980.
[15] L. J. Heyer, S. Kruglyak and S. Yooseph, “Exploring Ex-
pression Data: Identification and Analysis of Coexpressed
Genes,” Genome Research, Vol. 9, No. 11, 1999, pp.
1106-1115.
[16] C. C. Fung, K. W. Wong, H. Eren, R. Charlebois and H.
Crocker, “Modular Artificial Neural Network for Predic-
tion of Petrophysical Properties from Well Log Data,”
IEEE Transactions on Instrumentation & Measurement,
Vol. 46, No. 6, December 1997, pp. 1259-1263.
[17] D. Myers, K. W. Wong and C. C. Fung, “Self-organising
Maps Use for Intelligent Data Analysis,” Australian
Journal of Intelligent Information Processing Systems,
Vol. 6 No. 2, 2000, pp. 89-96.
[18] D. R. Hill, “A Vector Clustering Technique,” In: Samuel-
son, Ed., Mechanized Information Storage, Retrieval and
Dissemination North-Holland, Amsterdam, 1968.
[19] J. J. Rocchio, “Document Retrieval Systems — Optimi-
zation and Evaluation,” Doctoral Thesis, Harvard Univer-
sity, Boston, 1966.
[20] A. W. G. Salton and C. S. Yang, “A Vector Space Model
for Automatic Indexing,” Communication of ACM, Vol.
18, No. 11, 1975, pp. 613-620.
[21] O. Zamir, “Clustering Web Documents: A Phrase-Based
Method for Group Search Engine Results,” Computer
Science & Engineering, Ph.D. Thesis, University of
Washington, 1999.
[22] M. F. H. J. Bakus and M. Kamel, “A SOM-Based Docu-
ment Clustering Using Phrases,” Proceeding of the 9th
International Conference on Neural Information Proc-
essing (ICONIP’02), Vol. 5, 2002, pp. 2212-2216.
[23] D. Mladenic and M. Grobelnik, “Word Sequence as Fea-
tures in Text-learning,” Proceedings of the 17th Electro
technical and Computer Science Conference (ERK-98)
Ljubljana, Slovenia, 1998.
[24] K.-H. Tsai, C.-M. Tseng, C.-C. Hsu and H.-C. Chang, “On
the Chinese Document Clustering Based on Dynamical
Term Clustering,” Asia Information Retrieval Symposium
2005, Jeju Island, October 2005, pp. 534-539.
[25] C. Kruengkrai and C. Jaruskulchai, “Thai Text Document
Clustering Using Parallel Spherical K-means Algorithm
on PI-RUN Linux Cluster (in Thai),” The 5th National
Computer Science and Engineering Conference, Chiang
Mai University, Chiang Mai, 2001, pp. 7-9
[26] T. Kohonen, “Self-Organization and Associative Mem-
ory,” Springer Series in Information Sciences, Springer-
Verlag, Berlin, 1984, p. 125.
[27] T. Chumwatana, K. W. Wong and H. Xie “Frequent max
substring mining for indexing,” International Journal of
Computer Science and System Analysis (IJ CSSA ), India,
2008, pp. 179-184.
[28] T. Chumwatana, K. W. Wong and H. Xie “An Efficient
Text Mining Technique,” 9th Postgraduate Electrical En-
gineering & Computing Symposium (PEECS2008), Perth,
Australia, 2008, pp. 147-152.
[29] T. Chumwatana, K. W. Wong and H. Xie, “Using Fre-
quent Max Substring Technique for Thai Keyword Ex-
traction used in Thai Text Mining,” 2nd International
Conference on Soft Computing, Intelligent System and
Information Technology (ICSIIT 2010), Bali, 1-2 July
2010, pp. 309-314.
[30] T. Chumwatana, K. W. Wong and H. Xie, “Thai Text
Mining to Support Web Search for E-Commerce,” The
7th International Conference on e-Business 2008 (INCEB
2008), Bangkok, 2008, pp. 66-70.
[31] J. E. Hodges and Y. Wang, “Document Clustering using
Compound Words,” Proceedings of the 2005 Interna-
tional Conference on Artificial Intelligence (ICAI 2005),
Las Vegas, Nevada, 2005, pp. 307-313.