Rich document representation and classification: An analysis

There are three factors involved in text classification. These are classification model, similarity measure and document representation model. In this paper, we will focus on document representation and demonstrate that the choice of document representation has a profound impact on the quality of the classifier. In our experiments, we have used the centroid-based text classifier, which is a simple and robust text classification scheme. We will compare four different types of document representations: N-grams, Single terms, phrases and RDR which is a logic-based document representation. The N-gram representation is a string-based representation with no linguistic processing. The Single term approach is based on words with minimum linguistic processing. The phrase approach is based on linguistically formed phrases and single words. The RDR is based on linguistic processing and representing documents as a set of logical predicates. We have experimented with many text collections and we have obtained similar results. Here, we base our arguments on experiments conducted on Reuters-21578. We show that RDR, the more complex representation, produces more effective classifier on Reuters-21578, followed by the phrase approach.     

A segment-based approach to clustering multi-topic documents

Document clustering has been recognized as a central problem in text data management. Such a problem becomes particularly challenging when document contents are characterized by subtopical discussions that are not necessarily relevant to each other. Existing methods for document clustering have traditionally assumed that a document is an indivisible unit for text representation and similarity computation, which may not be appropriate to handle documents with multiple topics. In this paper, we address the problem of multi-topic document clustering by leveraging the natural composition of documents in text segments that are coherent with respect to the underlying subtopics. We propose a novel document clustering framework that is designed to induce a document organization from the identification of cohesive groups of segment-based portions of the original documents. We empirically give evidence of the significance of our segment-based approach on large collections of multi-topic documents, and we compare it to conventional methods for document clustering.     

Using wavelet analysis for text categorization in digital libraries: a first experiment with Strathprints

Digital libraries increasingly benefit from research on automated text categorization for improved access. Such research is typically carried out by means of standard test collections. In this article, we present a pilot experiment of replacing such test collections by a set of 6,000 objects from a real-world digital repository, indexed by Library of Congress Subject Headings, and test support vector machines in a supervised learning setting for their ability to reproduce the existing classification. To augment the standard approach, we introduce a combination of two novel elements: using functions for document content representation in Hilbert space, and adding extra semantics from lexical resources to the representation. Results suggest that wavelet-based kernels slightly outperformed traditional kernels on classification reconstruction from abstracts and vice versa from full-text documents, the latter outcome being due to word sense ambiguity. The practical implementation of our methodological framework enhances the analysis and representation of specific knowledge relevant to large-scale digital collections, in this case the thematic coverage of the collections. Representation of specific knowledge about digital collections is one of the basic elements of the persistent archives and the less studied one (compared to representations of digital objects and collections). Our research is an initial step in this direction developing further the methodological approach and demonstrating that text categorization can be applied to analyse the thematic coverage in digital repositories.     

Hybrid neural document clustering using guided self-organization and WordNet

Document clustering is text processing that groups documents with similar concepts. It's usually considered an unsupervised learning approach because there's no teacher to guide the training process, and topical information is often assumed to be unavailable. A guided approach to document clustering that integrates linguistic top-down knowledge from WordNet into text vector representations based on the extended significance vector weighting technique improves both classification accuracy and average quantization error. In our guided self-organization approach we integrate topical and semantic information from WordNet. Because a document-training set with preclassified information implies relationships between a word and its preference class, we propose a novel document vector representation approach to extract these relationships for document clustering. Furthermore, merging statistical methods, competitive neural models, and semantic relationships from symbolic Word-Net, our hybrid learning approach is robust and scales up to a real-world task of clustering 100,000 news documents.     

Term frequency with average term occurrences for textual information retrieval

In the context of information retrieval (IR) from text documents, the term weighting scheme (TWS) is a key component of the matching mechanism when using the vector space model. In this paper, we propose a new TWS that is based on computing the average term occurrences of terms in documents and it also uses a discriminative approach based on the document centroid vector to remove less significant weights from the documents. We call our approach Term Frequency With Average Term Occurrence (TF-ATO). An analysis of commonly used document collections shows that test collections are not fully judged as achieving that is expensive and maybe infeasible for large collections. A document collection being fully judged means that every document in the collection acts as a relevant document to a specific query or a group of queries. The discriminative approach used in our proposed approach is a heuristic method for improving the IR effectiveness and performance and it has the advantage of not requiring previous knowledge about relevance judgements. We compare the performance of the proposed TF-ATO to the well-known TF-IDF approach and show that using TF-ATO results in better effectiveness in both static and dynamic document collections. In addition, this paper investigates the impact that stop-words removal and our discriminative approach have on TF-IDF and TF-ATO. The results show that both, stop-words removal and the discriminative approach, have a positive effect on both term-weighting schemes. More importantly, it is shown that using the proposed discriminative approach is beneficial for improving IR effectiveness and performance with no information on the relevance judgement for the collection.     

Font adaptive word indexing of modern printed documents

We propose an approach for the word-level indexing of modern printed documents which are difficult to recognize using current OCR engines. By means of word-level indexing, it is possible to retrieve the position of words in a document, enabling queries involving proximity of terms. Web search engines implement this kind of indexing, allowing users to retrieve Web pages on the basis of their textual content. Nowadays, digital libraries hold collections of digitized documents that can be retrieved either by browsing the document images or relying on appropriate metadata assembled by domain experts. Word indexing tools would therefore increase the access to these collections. The proposed system is designed to index homogeneous document collections by automatically adapting to different languages and font styles without relying on OCR engines for character recognition. The approach is based on three main ideas: the use of self organizing maps (SOM) to perform unsupervised character clustering, the definition of one suitable vector-based word representation whose size depends on the word aspect-ratio, and the run-time alignment of the query word with indexed words to deal with broken and touching characters. The most appropriate applications are for processing modern printed documents (17th to 19th centuries) where current OCR engines are less accurate. Our experimental analysis addresses six data sets containing documents ranging from books of the 17th century to contemporary journals.     

一种基于新词发现的Web文本表示方法

Web文本表示是Web文本特征提取和分类的前提,最常用的文本表示是向量空间模型（VSM）,其中向量一般是基于词的特征项。由于向量空间模型本身没有考虑文本上下文间的潜在概念结构（如词汇间的共现关系）,而Web文本是一种半结构化文本,同时经常有新词出现,因此在VSM基础上提出了一种基于新词发现的Web文本表示方法：首先进行预处理将网页转化为文本;然后进行文本分词;接着通过二元互信息进行新词发现,同时把新词加入字典重新分词;最后用词和新词共同来表示Web文本。实验结果表明,该方法可以帮助识别未登录词并扩充现有字典,能够增强Web文本表示能力,改善Web文本的特征项质量,提高Web文本分类效果。     

一种大容量文本集的智能检索方法

分析了潜在语义模型,研究了潜在语义空间中文本的表示方法,提出了一种大容量文本集的检索策略。检索过程由粗粒度非相关剔除和相关文本的精确检索两个步骤组成。使用潜在语义空间模型对文本集进行初步的筛选,剔除非相关文本;使用大规模文本检索方法对相关文本在段落一级进行精确检索,其中为了提高检索的执行效率,在检索算法中引入了遗传算法;输出这些候选的段落序号。实验结果证明了这种方法的有效性和高效性。     

Use of text signatures for document retrieval in a highly parallel environment

This paper considers the use of text signatures, fixed-length bit string representations of document content, in an experimental information retrieval system: such signatures may be generated from the list of keywords characterising a document or a query. A file of documents may be searched in a bit-serial parallel computer, such as the ICL Distributed Array Processor, using a two-level retrieval strategy in which a comparison of a query signature with the file of document signatures provides a simple and efficient means of identifying those few documents that need to undergo a computationally demanding, character matching search. Text retrieval experiments using three large collections of documents and queries demonstrate the efficiency of the suggested approach.     

Fast Extraction of Semantic Features from a Latent Semantic Indexed Text Corpus

This paper proposes a projection-based symmetrical factorisation method for extracting semantic features from collections of text documents stored in a Latent Semantic space. Preliminary experimental results demonstrate this yields a comparable representation to that provided by a novel probabilistic approach which reconsiders the entire indexing problem of text documents and works directly in the original high dimensional vector-space representation of text. The employed projection index is derived here from the a priori constraints on the problem. The principal advantage of this approach is computational efficiency and is obtained by the exploitation of the Latent Semantic Indexing as a preprocessing stage. Simulation results on subsets of the 20-Newsgroups text corpus in various settings are provided. This revised version was published online in August 2006 with corrections to the Cover Date.     

LSISOM — A Latent Semantic Indexing Approach to Self-Organizing Maps of Document Collections

The Self Organizing Map (SOM) algorithm has been utilized, with much success, in a variety of applications for the automatic organization of full-text document collections. A great advantage of the SOM method is that document collections can be ordered in such a way so that documents with similar content are positioned at nearby locations of the 2-dimensional SOM lattice. The resulting ordered map thus presents a general view of the document collection which helps the exploration of information contained in the whole document space. The most notable example of such an application is the WEBSOM method where the document collection is ordered onto a map by utilizing word category histograms for representing the documents data vectors. In this paper, we introduce the LSISOM method which resembles WEBSOM in the sense that the document maps are generated from word category histograms rather than simple histograms of the words. However, a major difference between the two methods is that in WEBSOM the word category histograms are formed using statistical information of short word contexts whereas in LSISOM these histograms are obtained from the SOM clustering of the Latent Semantic Indexing representation of document terms.     

Conceptual feature generation for textual information using a conceptual network constructed from Wikipedia

A proper semantic representation of textual information underlies many natural language processing tasks. In this paper, a novel semantic annotator is presented to generate conceptual features for text documents. A comprehensive conceptual network is automatically constructed with the aid of Wikipedia that has been represented as a Markov chain. Furthermore, semantic annotator gets a fragment of natural language text and initiates a random walk to generate conceptual features that represent topical semantic of the input text. The generated conceptual features are applicable to many natural language processing tasks where the input is textual information and the output is a decision based on its context. Consequently, the effectiveness of the generated features is evaluated in the task of document clustering and classification. Empirical results demonstrate that representing text using conceptual features and considering the relations between concepts can significantly improve not only the bag of words representation but also other state‐of‐the‐art approaches.     

Unsupervised document structure analysis of digital scientific articles

Text mining and information retrieval in large collections of scientific literature require automated processing systems that analyse the documents’ content. However, the layout of scientific articles is highly varying across publishers, and common digital document formats are optimised for presentation, but lack structural information. To overcome these challenges, we have developed a processing pipeline that analyses the structure a PDF document using a number of unsupervised machine learning techniques and heuristics. Apart from the meta-data extraction, which we reused from previous work, our system uses only information available from the current document and does not require any pre-trained model. First, contiguous text blocks are extracted from the raw character stream. Next, we determine geometrical relations between these blocks, which, together with geometrical and font information, are then used categorize the blocks into different classes. Based on this resulting logical structure we finally extract the body text and the table of contents of a scientific article. We separately evaluate the individual stages of our pipeline on a number of different datasets and compare it with other document structure analysis approaches. We show that it outperforms a state-of-the-art system in terms of the quality of the extracted body text and table of contents. Our unsupervised approach could provide a basis for advanced digital library scenarios that involve diverse and dynamic corpora.     

Document Similarity Using a Phrase Indexing Graph Model

Document clustering techniques mostly rely on single term analysis of text, such as the vector space model. To better capture the structure of documents, the underlying data model should be able to represent the phrases in the document as well as single terms. We present a novel data model, the Document Index Graph, which indexes Web documents based on phrases rather than on single terms only. The semistructured Web documents help in identifying potential phrases that when matched with other documents indicate strong similarity between the documents. The Document Index Graph captures this information, and finding significant matching phrases between documents becomes easy and efficient with such model. The model is flexible in that it could revert to a compact representation of the vector space model if we choose not to index phrases. However, using phrase indexing yields more accurate document similarity calculations. The similarity between documents is based on both single term weights and matching phrase weights. The combined similarities are used with standard document clustering techniques to test their effect on the clustering quality. Experimental results show that our phrase-based similarity, combined with single-term similarity measures, gives a more accurate measure of document similarity and thus significantly enhances Web document clustering quality.     

Text Retrieval from Document Images Based on Word Shape Analysis

In this paper, we propose a method of text retrieval from document images using a similarity measure based on word shape analysis. We directly extract image features instead of using optical character recognition. Document images are segmented into word units and then features called vertical bar patterns are extracted from these word units through local extrema points detection. All vertical bar patterns are used to build document vectors. Lastly, we obtain the pair-wise similarity of document images by means of the scalar product of the document vectors. Four corpora of news articles were used to test the validity of our method. During the test, the similarity of document images using this method was compared with the result of ASCII version of those documents based on the N-gram algorithm for text documents.     

Query-driven indexing for scalable peer-to-peer text retrieval

In this paper, we present a query-driven indexing/retrieval strategy for efficient full text retrieval from large document collections distributed within a structured P2P network. Our indexing strategy is based on two important properties: (1) the generated distributed index stores posting lists for carefully chosen indexing term combinations that are frequently present in user queries, and (2) the posting lists containing too many document references are truncated to a bounded number of their top-ranked elements. These two properties guarantee acceptable latency and bandwidth requirements, essentially because the number of indexing term combinations remains scalable and the posting lists transmitted during retrieval never exceed a constant size. A novel index update mechanism efficiently handles adding of new documents to the document collection. Thus, the generated distributed index corresponds to a constantly evolving query-driven indexing structure that efficiently follows current information needs of the users and changes in the document collection.We show that the size of the index and the generated indexing/retrieval traffic remains manageable even for Web-size document collections at the price of a marginal loss in precision for rare queries. Our theoretical analysis and experimental results provide convincing evidence about the feasibility of the query-driven indexing strategy for large scale P2P text retrieval.     

Text document clustering using global term context vectors

Despite the advantages of the traditional vector space model (VSM) representation, there are known deficiencies concerning the term independence assumption. The high dimensionality and sparsity of the text feature space and phenomena such as polysemy and synonymy can only be handled if a way is provided to measure term similarity. Many approaches have been proposed that map document vectors onto a new feature space where learning algorithms can achieve better solutions. This paper presents the global term context vector-VSM (GTCV-VSM) method for text document representation. It is an extension to VSM that: (i) it captures local contextual information for each term occurrence in the term sequences of documents; (ii) the local contexts for the occurrences of a term are combined to define the global context of that term; (iii) using the global context of all terms a proper semantic matrix is constructed; (iv) this matrix is further used to linearly map traditional VSM (Bag of Words—BOW) document vectors onto a ‘semantically smoothed’ feature space where problems such as text document clustering can be solved more efficiently. We present an experimental study demonstrating the improvement of clustering results when the proposed GTCV-VSM representation is used compared with traditional VSM-based approaches.     

A hybrid approach for estimating document frequencies in unstructured P2P networks

Scalable search and retrieval over numerous web document collections distributed across different sites can be achieved by adopting a peer-to-peer (P2P) communication model. Terms and their document frequencies are the main components of text information retrieval and as such need to be computed, aggregated, and distributed throughout the system. This is a challenging problem in the context of unstructured P2P networks, since the local document collections may not reflect the global collection in an accurate way. This might happen due to skews in the distribution of documents to peers. Moreover, central assembly of the total information is not a scalable solution due to the excessive cost of storage and maintenance, and because of issues related to digital rights management. In this paper, we present an efficient hybrid approach for aggregation of document frequencies using a hierarchical overlay network for a carefully selected set of the most important terms, together with gossip-based aggregation for the remaining terms in the collections. Furthermore, we present a cost analysis to compute the communication cost of hybrid aggregation. We conduct experiments on three document collections, in order to evaluate the quality of the proposed hybrid aggregation.     

A soft image representation approach by exploiting local neighborhood structure of self-organizing map (SOM)

When images are described with visual words based on vector quantization of low-level color, texture, and edge-related visual features of image regions, it is usually referred as “bag-of-visual words (BoVW)”-based presentation. Although it has proved to be effective for image representation similar to document representation in text retrieval, the hard image encoding approach based on one-to-one mapping of regions to visual words is not expressive enough to characterize the image contents with higher level semantics and prone to quantization error. Each word is considered independent of all the words in this model. However, it is found that the words are related and their similarity of occurrence in documents can reflect the underlying semantic relations between them. To consider this, a soft image representation scheme is proposed by spreading each region’s membership values through a local fuzzy membership function in a neighborhood to all the words in a codebook generated by self-organizing map (SOM). The topology preserving property of the SOM map is exploited to generate a local membership function. A systematic evaluation of retrieval results of the proposed soft representation on two different image (natural photographic and medical) collections has shown significant improvement in precision at different recall levels when compared to different low-level and “BoVW”-based feature that consider only probability of occurrence (or presence/absence) of a word.