Latent Semantic Analysis and Beyond

Latent Semantic Analysis and Beyond

Anne Kao (The Boeing Phantom Works, USA)
Copyright: © 2009 |Pages: 25
DOI: 10.4018/978-1-59904-990-8.ch032
OnDemand PDF Download:


Latent Semantic Analysis (LSA) or Latent Semantic Indexing (LSI), when applied to information retrieval, has been a major analysis approach in text mining. It is an extension of the vector space method in information retrieval, representing documents as numerical vectors but using a more sophisticated mathematical approach to characterize the essential features of the documents and reduce the number of features in the search space. This chapter summarizes several major approaches to this dimensionality reduction, each of which has strengths and weaknesses, and it describes recent breakthroughs and advances. It shows how the constructs and products of LSA applications can be made user-interpretable and reviews applications of LSA beyond information retrieval, in particular, to text information visualization.
Chapter Preview


A vast amount of information exists in text form, such as free (unstructured) or semi-structured text, including many database fields, reports, memos, e-mail, Web sites, and news articles. Various Web mining and text mining methods have been developed to analyze textual resources. Latent Semantic Analysis (LSA) (Deerwester, Dumais, Furnas, Landauer, & Harshman, 1990), or Latent Semantic Indexing (LSI) when it is applied to document retrieval, has been a major approach in text mining. It is an extension of the vector space method in Information Retrieval (Salton, Wong, & Yang, 1975), using a mathematical approach to represent documents as numerical vectors but with a more sophisticated means of characterizing the essential features of documents and reducing the number of dimensions needed to describe documents to a manageable size. There have been several major approaches to address this dimensionality reduction, each of which has strengths and weaknesses. A major challenge in using LSA is that it is typically considered a black box approach that makes it difficult to understand or interpret the results. However, more recent research has not only overcome this challenge, but also demonstrates that the use of LSA extends beyond IR and text document clustering to become a major player in the area of text information visualization. This chapter will summarize the major approaches to LSA, their strengths and weakness, as well as recent breakthroughs and advances and applications beyond information retrieval.

Text mining has adopted certain techniques from the more general field of data analysis, including sophisticated methods for analyzing relationships among highly formatted data, such as numerical data or data with a relatively small fixed number of possible values. Such techniques can expose patterns and trends in this type of data. Text mining can identify relationships between individual unstructured or semi-structured text documents, as well as more general semantic patterns across large collections of such documents. Latent Semantic Analysis, like many other methods of text mining, depends on the twin concepts of “document” and “term.” As used in this chapter, a “document” refers to any body of unstructured or semi-structured text. The text may include the entire content of a document in the general sense, such as a book, an article, a paper, or the like—or only a portion of a document, such as an abstract, a paragraph, a sentence, or a title. Ideally, a “document” describes a coherent topic. In addition, a “document” can be the text field of a database, or encompass text generated from an image or graphic, or it may be text recovered from audio or video formats. We will use the term “document” in this general sense.

A document can be represented as a collection of “terms,” each of which can appear in multiple documents. Typically, a “term” consists of an individual word used in the text. However, a “term” can also include multiple words that are commonly used together, for example, “landing gear”, or even consist of a string that need not appear explicitly in the text but rather result from token normalization or standardization. Token normalization will be discussed further.

In vector-based methods of text data analysis, after a suitable set of terms has been defined for a document collection, the collection can be represented as a set of vectors. With traditional vector space methods, individual documents are treated as vectors in a high-dimensional vector space in which each dimension corresponds to some feature of a document, typically a term. A collection of documents can thus be represented by a two-dimensional matrix A(t,d) of features (terms) and documents. In the typical case, the value of each matrix entry is the number of occurrences of that term in the specified document, or some weighting or principled transformation of that number. LSA, as an extension of the vector space method, involves methods of transforming A by various means, e.g. singular value decomposition (SVD) in the case of ‘classical’ LSA, which typically attempt to provide a more sophisticated set of features that better capture the latent semantics of the documents. We discuss various such matrix decomposition techniques in much more detail.

Key Terms in this Chapter

Latent Semantic Space: The subspace of term space whose dimensions correspond to the features uncovered by Latent Semantic Analysis for representing documents in a more semantically useful form.

Singular Value Decomposition (SVD): A linear algebra method of decomposing an arbitrary matrix into three matrices, two of which are orthonormal (the columns, the left and right singular vectors, respectively, are orthogonal and have length1) and the third is a diagonal matrix whose diagonal values are the singular values of the matrix

Basis Vectors (for a given space): A set of linearly independent vectors that define a space in that any vector in that space can be defined as a linear combination (a weighted sum) of those vectors. Linearly independent means that none of them can be defined as a linear combination (or weighted sum) of the others.

Latent Semantic Analysis (LSA): A method of representing text documents in terms of features that are weighted combinations of the frequencies words or terms in the documents that makes the “latent semantics” or topics treated in the documents more computationally accessible.

Subspace: A vector space with a lower dimensionality that is wholly contained in a larger vector space.

Dimensionality Reduction: The process of taking high dimensional data (data represented by a large number of features) and representing it with different and fewer features or dimensions (which may be combinations of the old features) in a principled fashion that preserves some properties of the original space.

Vector Space Methods: A method of representing documents as numerical vectors, where the values represent the frequencies of the words or terms in the documents, or some weighting of these to represent their importance in the document set

Principal Components Analysis (PCA): A statistical method for discovering the dimensions that maximize variability in high dimensional data. Mathematically equivalent to SVD, except that it requires that the data all be centered

Top ic Words: Words that summarize the important topics of a document or piece of text that are automatically assigned based on the representation of that document in latent semantic space.

Complete Chapter List

Search this Book:
Editorial Advisory Board
Table of Contents
Xiaohua Hu
Min Song, Yi-Fang Brook Wu
Min Song, Yi-Fang Brook Wu
Chapter 1
Ying Liu
In the automated text classification, a bag-of-words representation followed by the tfidf weighting is the most popular approach to convert the... Sample PDF
On Document Representation and Term Weights in Text Classification
Chapter 2
Yi-fang Brook Wu, Quanzhi Li
Document keyphrases provide semantic metadata which can characterize documents and produce an overview of the content of a document. This chapter... Sample PDF
Deriving Document Keyphrases for Text Mining
Chapter 3
John Atkinson
This chapter introduces a novel evolutionary model for intelligent text mining. The model deals with issues concerning shallow text representation... Sample PDF
Intelligent Text Mining: Putting Evolutionary Methods and Language Technologies Together
Chapter 4
Xiaoyan Yu, Manas Tungare, Weigo Yuan, Yubo Yuan, Manuel Pérez-Quiñones, Edward A. Fox
Syllabi are important educational resources. Gathering syllabi that are freely available and creating useful services on top of the collection... Sample PDF
Automatic Syllabus Classification Using Support Vector Machines
Chapter 5
Xiao-Li Li
In traditional text categorization, a classifier is built using labeled training documents from a set of predefined classes. This chapter studies a... Sample PDF
Partially Supervised Text Categorization
Chapter 6
Yu-Jin Zhang
Mining techniques can play an important role in automatic image classification and content-based retrieval. A novel method for image classification... Sample PDF
Image Classification and Retrieval with Mining Technologies
Chapter 7
Han-joon Kim
This chapter introduces two practical techniques for improving Naïve Bayes text classifiers that are widely used for text classification. The Naïve... Sample PDF
Improving Techniques for Naïve Bayes Text Classifiers
Chapter 8
Ricco Rakotomalala, Faouzi Mhamdi
In this chapter, we are interested in proteins classification starting from their primary structures. The goal is to automatically affect proteins... Sample PDF
Using the Text Categorization Framework for Protein Classification
Chapter 9
Wilson Wong
Feature-based semantic measurements have played a dominant role in conventional data clustering algorithms for many existing applications. However... Sample PDF
Featureless Data Clustering
Chapter 10
Xiaohui Cui
In this chapter, we introduce three nature inspired swarm intelligence clustering approaches for document clustering analysis. The major challenge... Sample PDF
Swarm Intelligence in Text Document Clustering
Chapter 11
P. Viswanth
Clustering is a process of finding natural grouping present in a dataset. Various clustering methods are proposed to work with various types of... Sample PDF
Some Efficient and Fast Approaches to Document Clustering
Chapter 12
Abdelmalek Amine, Zakaria Elberrichi, Michel Simonet, Ladjel Bellatreche, Mimoun Malki
The classification of textual documents has been the subject of many studies. Technologies like the Web and numerical libraries facilitated the... Sample PDF
SOM-Based Clustering of Textual Documents Using WordNet
Chapter 13
Lean Yu, Shouyang Wang, Kin Keung Lai
With the rapid increase of the huge amount of online information, there is a strong demand for Web text mining which helps people discover some... Sample PDF
A Multi-Agent Neural Network System for Web Text Mining
Chapter 14
Sangeetha Kutty
With the emergence of XML standardization, XML documents have been widely used and accepted in almost all the major industries. As a result of the... Sample PDF
Frequent Mining on XML Documents
Chapter 15
Richi Nayak
XML has gained popularity for information representation, exchange and retrieval. As XML material becomes more abundant, its heterogeneity and... Sample PDF
The Process and Application of XML Data Mining
Chapter 16
Francesco Buccafurri
In the context of Knowledge Discovery in Databases, data reduction is a pre-processing step delivering succinct yet meaningful data to sequent... Sample PDF
Approximate Range Querying over Sliding Windows
Chapter 17
Jan H. Kroeze
This chapter discusses the application of some data warehousing techniques on a data cube of linguistic data. The results of various modules of... Sample PDF
Slicing and Dicing a Linguistic Data Cube
Chapter 18
Yi-fang Brook Wu, Xin Chen
This chapter presents a methodology for personalized knowledge discovery from text. Traditionally, problems with text mining are numerous rules... Sample PDF
Discovering Personalized Novel Knowledge from Text
Chapter 19
Catia Pesquita
Biomedical research generates a vast amount of information that is ultimately stored in scientific publications or in databases. The information in... Sample PDF
Untangling BioOntologies for Mining Biomedical Information
Chapter 20
Luis M. de Campos
In this chapter, we present a thesaurus application in the field of text mining and more specifically automatic indexing on the set of descriptors... Sample PDF
Thesaurus-Based Automatic Indexing
Chapter 21
Concept-Based Text Mining  (pages 346-358)
Stanley Loh, Leandro Krug Wives, Daniel Lichtnow, José Palazzo M. de Oliveira
The goal of this chapter is to present an approach to mine texts through the analysis of higher level characteristics (called “concepts’)... Sample PDF
Concept-Based Text Mining
Chapter 22
Marcello Pecoraro
This chapter aims at providing an overview about the use of statistical methods supporting the Web Usage Mining. Within the first part is described... Sample PDF
Statistical Methods for User Profiling in Web Usage Mining
Chapter 23
Quanzhi Li, Yi-fang Brook Wu
This chapter presents a new approach of mining the Web to identify people of similar background. To find similar people from the Web for a given... Sample PDF
Web Mining to Identify People of Similar Background
Chapter 24
Pawan Lingras
This chapter describes how Web usage patterns can be used to improve the navigational structure of a Web site. The discussion begins with an... Sample PDF
Hyperlink Structure Inspired by Web Usage
Chapter 25
Rosa Meo, Maristella Matera
In this chapter, we present the usage of a modeling language, WebML, for the design and the management of dynamic Web applications. WebML also makes... Sample PDF
Designing and Mining Web Applications: A Conceptual Modeling Approach
Chapter 26
Brigitte Trousse, Marie-Aude Aufaure, Bénédicte Le Grand, Yves Lechevallier, Florent Masseglia
This chapter proposes an original approach for ontology management in the context of Web-based information systems. Our approach relies on the usage... Sample PDF
Web Usage Mining for Ontology Management
Chapter 27
Yue-Shi Lee
Web mining is one of the mining technologies, which applies data mining techniques in large amounts of Web data to improve the Web services. Web... Sample PDF
A Lattice-Based Framework for Interactively and Incrementally Mining Web Traversal Patterns
Chapter 28
Stanley R.M. Oliveira, Osmar R. Zaïane
Privacy-preserving data mining (PPDM) is one of the newest trends in privacy and security research. It is driven by one of the major policy issues... Sample PDF
Privacy-Preserving Data Mining on the Web: Foundations and Techniques
Chapter 29
G.S. Mahalakshmi, S. Sendhilkumar
Automatic reference tracking involves systematic tracking of reference articles listed for a particular research paper by extracting the references... Sample PDF
Automatic Reference Tracking
Chapter 30
Wilson Wong
As more electronic text is readily available, and more applications become knowledge intensive and ontology-enabled, term extraction, also known as... Sample PDF
Determination of Unithood and Termhood for Term Recognition
Chapter 31
Fotis Lazarinis
Over 60% of the online population are non-English speakers and it is probable the number of non-English speakers is growing faster than English... Sample PDF
Retrieving Non-Latin Information in a Latin Web: The Case of Greek
Chapter 32
Anne Kao
Latent Semantic Analysis (LSA) or Latent Semantic Indexing (LSI), when applied to information retrieval, has been a major analysis approach in text... Sample PDF
Latent Semantic Analysis and Beyond
Chapter 33
Ganesh Ramakrishnan, Pushpak Bhattacharyya
Text mining systems such as categorizers and query retrievers of the first generation were largely hinged on word level statistics and provided a... Sample PDF
Question Answering Using Word Associations
Chapter 34
Giuseppe Manco, Riccardo Ortale, Andrea Tagarelli
Personalization is aimed at adapting content delivery to users’ profiles: namely, their expectations, preferences and requirements. This chapter... Sample PDF
The Scent of a Newsgroup: Providing Personalized Access to Usenet Sites through Web Mining
Chapter 35
Alexander Dreweke, Ingrid Fischer, Tobias Werth, Marc Wörlein
Searching for frequent pieces in a database with some sort of text is a well-known problem. A special sort of text is program code as e.g. C++ or... Sample PDF
Text Mining in Program Code
Chapter 36
Nitin Agarwal, Huan Liu, Jianping Zhang
In Golbeck and Hendler (2006), authors consider those social friendship networking sites where users explicitly provide trust ratings to other... Sample PDF
A Study of Friendship Networks and Blogosphere
Chapter 37
Pasquale De Meo
In this chapter we present an information system conceived for supporting managers of Public Health Care Agencies to decide the new health care... Sample PDF
An HL7-Aware Decision Support System for E-Health
Chapter 38
Diego Liberati
Building effective multitarget classifiers is still an on-going research issue: this chapter proposes the use of the knowledge gleaned from a human... Sample PDF
Multitarget Classifiers for Mining in Bioinformatics
Chapter 39
Shuting Xu
Text mining is an instrumental technology that today’s organizations can employ to extract information and further evolve and create valuable... Sample PDF
Current Issues and Future Analysis in Text Mining for Information Security Applications
Chapter 40
E. Thirumaran
This chapter introduces Collaborative filtering-based recommendation systems, which has become an integral part of E-commerce applications, as can... Sample PDF
Collaborative Filtering Based Recommendation Systems
Chapter 41
Hanna Suominen
The purpose of this chapter is to provide an overview of prevalent measures for evaluating the quality of system output in seven key text mining... Sample PDF
Performance Evaluation Measures for Text Mining
Chapter 42
Yanliang Qi
The biology literatures have been increased in an exponential growth in recent year. The researchers need an effective tool to help them find out... Sample PDF
Text Mining in Bioinformatics: Research and Application
Chapter 43
Ki Jung Lee
With the increased use of Internet, a large number of consumers first consult on line resources for their healthcare decisions. The problem of the... Sample PDF
Literature Review in Computational Linguistics Issues in the Developing Field of Consumer Informatics: Finding the Right Information for Consumer's Health Information Need
Chapter 44
Richard S. Segall
This chapter presents background on text mining, and comparisons and summaries of seven selected software for text mining. The text mining software... Sample PDF
A Survey of Selected Software Technologies for Text Mining
Chapter 45
Ah Chung Tsoi, Phuong Kim To, Markus Hagenbuchner
This chapter describes the application of a number of text mining techniques to discover patterns in the health insurance schedule with an aim to... Sample PDF
Application of Text Mining Methodologies to Health Insurance Schedules
Chapter 46
Miao-Ling Wang, Hsiao-Fan Wang
With the ever-increasing and ever-changing flow of information available on the Web, information analysis has never been more important. Web text... Sample PDF
Web Mining System for Mobile-Phone Marketing
Chapter 47
Neil Davis
Text mining technology can be used to assist in finding relevant or novel information in large volumes of unstructured data, such as that which is... Sample PDF
Web Service Architectures for Text Mining: An Exploration of the Issues via an E-Science Demonstrator
About the Editors
About the Contributors