A two-stage feature selection method for text categorization

Feature selection for text categorization is a well-studied problem and its goal is to improve the effectiveness of categorization, or the efficiency of computation, or both. The system of text categorization based on traditional term-matching is used to represent the vector space model as a document; however, it needs a high dimensional space to represent the document, and does not take into account the semantic relationship between terms, which leads to a poor categorization accuracy. The latent semantic indexing method can overcome this problem by using statistically derived conceptual indices to replace the individual terms. With the purpose of improving the accuracy and efficiency of categorization, in this paper we propose a two-stage feature selection method. Firstly, we apply a novel feature selection method to reduce the dimension of terms; and then we construct a new semantic space, between terms, based on the latent semantic indexing method. Through some applications involving the spam database categorization, we find that our two-stage feature selection method performs better.     

基于语义的VSM模型改进

文本聚类在很多领域都有广泛的应用,传统的文本聚类方法由于并不考虑语义因素,得出的聚类效果并不理想.利用语义对VSM模型进行变换,即基于语义对VSM模型的各维进行扭曲,将原本的正交坐标系基于语义变换为斜角坐标系,然后将文本的特征向量映射到变换后的VSM模型上再进行聚类,相对减小语义相关的特征向量间的语义距离,从而提高了文...     

2DVTE: A two-directional videotext extractor for rapid and elaborate design

In video indexing and summarization, videotext is the very compact and accurate information. Most videotext detection and extraction methods only deal with the static videotext on video frames. Few methods can handle motion videotext efficiently since motion videotext is hardly extracted well. In this paper, we propose a two-directional videotext extractor, called 2DVTE. It is developed as an integrated system to detect, localize and extract the scrolling videotexts. First, the detection method is carried out by edge information to classify regions into text and non-text regions. Second, referring to the localization on scrolling videotext, we propose the two-dimensional projection profile method with horizontal and vertical edge map information. Considering the characteristics of Chinese text, the vertical edge map is used to localize the possible text region and horizontal edge map is used to refine the text region. Third, the extraction method consists of dual mode adaptive thresholding and multi-seed filling algorithm. In the dual mode adaptive thresholding, it produces the non-rectangle pattern to divide the background and foreground more precisely. Referring to the multi-seed filling algorithm, it is based on the consideration of the minimum and maximum length and four directions of the stroke while the previous method only considers the minimum length and two directions of the stroke. With this multi-seed exploitation on strokes, precise seeds are obtained to produce more sophisticated videotext. Considering high throughput and the low complexity issue, we can achieve a real-time system on detecting, localizing, and extracting the scrolling videotexts with only one frame usage instead of multi-frame integration in other literatures. According to the experiment results on various video sequences, all of the horizontal and vertical scrolling videotexts can be extracted precisely. We also make comparisons with other methods. In our analysis, the performance of our algorithm is superior to other existing methods in speed and quality.     

Text line and word segmentation of handwritten documents

In this paper, we present a segmentation methodology of handwritten documents in their distinct entities, namely, text lines and words. Text line segmentation is achieved by applying Hough transform on a subset of the document image connected components. A post-processing step includes the correction of possible false alarms, the detection of text lines that Hough transform failed to create and finally the efficient separation of vertically connected characters using a novel method based on skeletonization. Word segmentation is addressed as a two class problem. The distances between adjacent overlapped components in a text line are calculated using the combination of two distance metrics and each of them is categorized either as an inter- or an intra-word distance in a Gaussian mixture modeling framework. The performance of the proposed methodology is based on a consistent and concrete evaluation methodology that uses suitable performance measures in order to compare the text line segmentation and word segmentation results against the corresponding ground truth annotation. The efficiency of the proposed methodology is demonstrated by experimentation conducted on two different datasets: (a) on the test set of the ICDAR2007 handwriting segmentation competition and (b) on a set of historical handwritten documents.     

Discrete data clustering using finite mixture models

Finite mixture models have been applied for different computer vision, image processing and pattern recognition tasks. The majority of the work done concerning finite mixture models has focused on mixtures for continuous data. However, many applications involve and generate discrete data for which discrete mixtures are better suited. In this paper, we investigate the problem of discrete data modeling using finite mixture models. We propose a novel, well motivated mixture that we call the multinomial generalized Dirichlet mixture. The novel model is compared with other discrete mixtures. We designed experiments involving spatial color image databases modeling and summarization, and text classification to show the robustness, flexibility and merits of our approach.     

Genetic algorithm for text clustering based on latent semantic indexing

In this paper, we develop a genetic algorithm method based on a latent semantic model (GAL) for text clustering. The main difficulty in the application of genetic algorithms (GAs) for document clustering is thousands or even tens of thousands of dimensions in feature space which is typical for textual data. Because the most straightforward and popular approach represents texts with the vector space model (VSM), that is, each unique term in the vocabulary represents one dimension. Latent semantic indexing (LSI) is a successful technology in information retrieval which attempts to explore the latent semantics implied by a query or a document through representing them in a dimension-reduced space. Meanwhile, LSI takes into account the effects of synonymy and polysemy, which constructs a semantic structure in textual data. GA belongs to search techniques that can efficiently evolve the optimal solution in the reduced space. We propose a variable string length genetic algorithm which has been exploited for automatically evolving the proper number of clusters as well as providing near optimal data set clustering. GA can be used in conjunction with the reduced latent semantic structure and improve clustering efficiency and accuracy. The superiority of GAL approach over conventional GA applied in VSM model is demonstrated by providing good Reuter document clustering results.     

Diversity of Agents and Their Interaction

Diversity of agents occurs naturally in epistemic logic, and dynamic logics of information update and belief revision. In this paper we provide a systematic discussion of different sources of diversity, such as introspection ability, powers of observation, memory capacity, and revision policies, and we show how these can be encoded in dynamic epistemic logics allowing for individual variation among agents. Next, we explore the interaction of diverse agents by looking at some concrete scenarios of communication and learning, and we propose a logical methodology to deal with these as well. We conclude with some further questions on the logic of diversity and interaction. This work was supported by the Chinese National Social Science Foundation (Grant Number: 04CZX011) and the Dutch Science Organization NWO.     

Least squares twin support vector machines for pattern classification

In this paper we formulate a least squares version of the recently proposed twin support vector machine (TSVM) for binary classification. This formulation leads to extremely simple and fast algorithm for generating binary classifiers based on two non-parallel hyperplanes. Here we attempt to solve two modified primal problems of TSVM, instead of two dual problems usually solved. We show that the solution of the two modified primal problems reduces to solving just two systems of linear equations as opposed to solving two quadratic programming problems along with two systems of linear equations in TSVM. Classification using nonlinear kernel also leads to systems of linear equations. Our experiments on publicly available datasets indicate that the proposed least squares TSVM has comparable classification accuracy to that of TSVM but with considerably lesser computational time. Since linear least squares TSVM can easily handle large datasets, we further went on to investigate its efficiency for text categorization applications. Computational results demonstrate the effectiveness of the proposed method over linear proximal SVM on all the text corpuses considered.