Time filter

Source Type

Beijing, China

Beijing Language and Culture University , colloquially known in Chinese as Yuyan Xueyuan , has the main aim of teaching the Chinese language and culture to foreign students. However, it also takes Chinese students specializing in foreign languages and other relevant subjects of humanities and social science, and trains teachers of Chinese as a foreign language. It used to be the only institute of this kind in China. After the push for massification of higher education starting in the 90's, nowadays many other universities in almost every major city in China have a similar offer. Thus bachelor, master or post-doc degrees in, "Teaching Chinese to Foreigners", as well as bachelor and master degrees in several foreign languages are no longer only to be found at BLCU. Beijing Language and Culture University is often called "Little United Nations" in China because of its very large amount of international students from various countries. Wikipedia.

Liu G.,Beijing Language and Culture University
Proceedings - 2011 IEEE International Conference on Granular Computing, GrC 2011 | Year: 2011

Corresponding to two equivalent forms of Pawlak rough approximations, two different generalized rough sets, based on arbitrary relations, are proposed. This paper studies these two generalized rough sets. The relationship and difference between these two generalized rough sets are considered. © 2011 IEEE. Source

Liu Y.,Beijing Normal University | Hao M.,Beijing Language and Culture University | Li P.,Pennsylvania State University | Shu H.,Beijing Normal University
PLoS ONE | Year: 2011

The present study reports timed norms for 435 object pictures in Mandarin Chinese. These data include naming latency, name agreement, concept agreement, word length, and age of acquisition (AoA) based on children's naming and adult ratings, and several other adult ratings of concept familiarity, subjective word frequency, image agreement, image variability, and visual complexity. Furthermore, we examined factors that influence the naming latencies of the pictures. The results show that concept familiarity, AoA, concept agreement, name agreement, and image agreement are significant predictors of naming latencies, whereas subjective word frequency is not a reliable determinant. These results are discussed in light of picture naming data in other languages. An item-based index for the norms is provided in the Table S1. © 2011 Liu et al. Source

Liu G.,Beijing Language and Culture University | Sai Y.,Shandong University of Finance and Economics
Information Sciences | Year: 2010

This paper studies the classes of rough sets and fuzzy rough sets. We discuss the invertible lower and upper approximations and present the necessary and sufficient conditions for the lower approximation to coincide with the upper approximation in both rough sets and fuzzy rough sets. We also study the mathematical properties of a fuzzy rough set induced by a cyclic fuzzy relation. © 2010 Elsevier Inc. All rights reserved. Source

Xu Y.,Beijing Language and Culture University
Proceedings - 2011 IEEE International Conference on Granular Computing, GrC 2011 | Year: 2011

Feature selection plays an important role in text categorization. Automatic feature selection methods such as document frequency thresholds (DF), information gain (IG), mutual information (MI), and so on are commonly applied in text categorization. Existing experiments show IG is one of the most effective methods. In this paper, a feature selection method is proposed based on Rough Set theory and according to Rough set theory, knowledge about a universe of objects may be defined as classifications based on certain properties of the objects, i.e. rough set theory assume that knowledge is an ability to partition objects. We quantify the ability of classify objects, and call the amount of this ability as knowledge quantity and then following this quantification we put forward a notion "knowledge Gain" and propose a knowledge gain feature selection method (KG method)The task of spam filtering can be seen as a special problem of text classification. An effective and efficient feature selection method is important, which can be easily and effectively select the major features to attain the goal for anti-spam filtering. We explore 2 classifiers (Naive Bayes and SVM), and run our experiments on Chinese-spam collection show that KG performs better than the IG method, specially, on extremely aggressive reduction. We conclude that the KG feature method have a state-of-the-art performance for filtering spam, especially for Chinese spam emails. © 2011 IEEE. Source

Liu G.,Beijing Language and Culture University
Fundamenta Informaticae | Year: 2015

This paper studies quasi-discrete closure spaces and fuzzy closure spaces. We show that any topological closure cT induced by a closure c is the smallest extension from a closure space to a topological closure space in both crisp and fuzzy environment, in addition, a characterization of the continuous mappings in quasi-discrete closure spaces is obtained. We propose the concept of quasi-discrete fuzzy closure spaces in the context of fuzzy sets and establish a one to one correspondence between quasi-discrete fuzzy closure spaces and reflexive fuzzy relations. We also discuss the relationship between topological closure cT and closure c in quasi-discrete fuzzy closure spaces and show that the process from closure c to topological closure cT can be realized via the process from a reflexive fuzzy relation to its transitive closure. Source

Discover hidden collaborations