Feature selection for unsupervised tasks is particularly challenging, especially when dealing with text data. The increase in online documents and email communication creates a need for tools that can operate without the supervision of the user. In this paper we look at novel feature selection techniques that address this need. A distributional similarity measure from information theory is appl...

Feature selection is vital in the field of pattern classification due to accuracy and processing time considerations. The selection of proper features is of greater importance when the initial feature set is considerably large. Text classification is a typical example of this situation, where the size of the initial feature set may reach to hundreds or even thousands. There are numerous researc...

Very large corpora are increasingly exploited to improve Natural Language Processing (NLP) Systems. This however implies that the lexical, morpho−syntactic and syntactic homogeneity of the data used are mastered. This control in turn requires the development of tools aimed at text calibration or profiling. We are implementing such profiling tools and developing an associated methodology within ...

The Problem: All classifiers require a set of features that can be used to distinguish between different examples. In some cases, such as determining whether a chess position is a winning position, the features are clear (the positions of the chess pieces). In other cases, such as text, they are less clear. A document is simply a string of characters. Standard practice dictates that documents s...

Arbitrary-shaped scene text detection is a challenging task due to the variety of changes in font, size, color, and orientation. Most existing regression based methods resort regress masks or contour points regions model instances. However, regressing complete requires high training complexity, are not sufficient capture details highly curved texts. To tackle above limitations, we propose novel...

Feature selection method for text classification based on information gain ranking, improved by removing redundant terms using mutual information measure and inclusion index, is proposed. We report an experiment to study the impact of term redundancy on the performance of text classifier. The result shows that term redundancy behaves very similar to noise and may degrade the classifier performa...

Social Network Sites (SNSs) offer a plethora of privacy controls, but users rarely exploit all of these mechanisms, nor do they do so in the same manner. We demonstrate that SNS users instead adhere to one of a small set of distinct privacy management strategies that are partially related to their level of privacy feature awareness. Using advanced Factor Analysis methods on the self-reported pr...