基于主动学习的微博聚类分析

K Means聚类算法由于无法准确确定初始化聚类中心,容易造成 聚类结果准确率低下。对微博数据聚类时,可能会导致无法正确反映兴趣热点。本文 设计了基于主动学习的聚类算法,在确定初始聚类中心过程中应用Min Max主动学习策略, 使 得算法每次在很小数量的查询后都会提供数据点供用户进行初始中心点确认,并在K Means算 法中重新计算聚类中心时设置其权重值,从而减少迭代的数量,提高聚类结果的准确 率,并将这一算法运用于微博聚类分析,得出微博热门话题。     

Rotation,scaling and translation invariant texture recognition by Bessel-Fourier moments

The ideal of Bessel-Fourier moments (BFMs) for image analysis and only rotation invariant image cognition has been proposed recently. In this paper, we extend the previous work and propose a new method for rotation, scaling and translation (RST) invariant texture recognition using Bessel-Fourier moments. Compared with the others moments based methods, the radial polynomials of Bessel-Fourier moments have more zeros and these zeros are more evenly distributed. It makes Bessel-Fourier moments more suitable for invariant texture recognition as a generalization of orthogonal complex moments. In the experiment part, we got three testing sets of 16, 24 and 54 texture images by way of translating, rotating and scaling them separately. The correct classification percentages (CCPs) are compared with that of orthogonal Fourier-Mellin moments and Zernike moments based methods in both noise-free and noisy condition. Experimental results validate the conclusion of theoretical derivation: BFM performs better in recognition capability and noise robustness in terms of RST texture recognition under both noise-free and noisy condition when compared with orthogonal Fourier-Mellin moments and Zernike moments based methods.     

Semi‐global and global containment control of multi‐agent systems with second‐order dynamics and input saturation

This paper considers both semi‐global and global containment control for a second‐order multi‐agent system that is composed by a network of identical harmonic oscillators or double integrators with multiple leaders and input saturation. A distributed low gain feedback algorithm is proposed to solve the semi‐global containment control problem for the network whose topology is directed and initial condition is taken from any a priori given bounded set. In particular, by using a parametric Lyapunov equation approach, M‐matrix properties and algebraic graph theory, an upper bound of the low gain parameter is estimated such that the low gain feedback matrix can be analytically determined without involving numerical computation. Furthermore, under the assumption that the induced subgraph formed by the followers is strongly connected and detail balanced, two linear feedback protocols are designed for coupled harmonic oscillators and coupled double integrators, respectively, to asymptotically achieve the global containment control of the network with any initial condition. Finally, numerical examples are given to illustrate the effectiveness of the theoretical results. Copyright © 2016 John Wiley & Sons, Ltd.     

Combination of bottleneck feature extraction and dereverberation for distant-talking speech recognition

The performance of speech recognition in distant-talking environments is severely degraded by the reverberation that can occur in enclosed spaces (e.g., meeting rooms). To mitigate this degradation, dereverberation techniques such as network structure-based denoising autoencoders and multi-step linear prediction are used to improve the recognition accuracy of reverberant speech. Regardless of the reverberant conditions, a novel discriminative bottleneck feature extraction approach has been demonstrated to be effective for speech recognition under a range of conditions. As bottleneck feature extraction is not primarily designed for dereverberation, we are interested in whether it can compensate for other carefully designed dereverberation approaches. In this paper, we propose three schemes covering both front-end processing (cascaded combination and parallel combination) and back-end processing (system combination). Each of these schemes integrates bottleneck feature extraction with dereverberation. The effectiveness of these schemes is evaluated via a series of experiments using the REVERB challenge dataset.     

Arithmetic coding using hierarchical dependency context model for H.264/AVC video coding

In this paper, a hierarchical dependency context model (HDCM) is firstly proposed to exploit the statistical correlations of DCT (Discrete Cosine Transform) coefficients in H.264/AVC video coding standard, in which the number of non-zero coefficients in a DCT block and the scanned position are used to capture the magnitude varying tendency of DCT coefficients. Then a new binary arithmetic coding using hierarchical dependency context model (HDCMBAC) is proposed. HDCMBAC associates HDCM with binary arithmetic coding to code the syntax elements for a DCT block, which consist of the number of non-zero coefficients, significant flag and level information. Experimental results demonstrate that HDCMBAC can achieve similar coding performance as CABAC at low and high QPs (quantization parameter). Meanwhile the context modeling and the arithmetic decoding in HDCMBAC can be carried out in parallel, since the context dependency only exists among different parts of basic syntax elements in HDCM.     

Towards understanding the gamification upon users’ scores in a location-based social network

Online social platform, such as Wikipedia and Foursquare, has been increasingly exploded due to not only various useful services provided but also social gaming mechanisms that can keep users actively engaged. For example, users are awarded ”virtual goods” like badges and points when they contribute to the community in the network by voluntarily sharing ideas and other information. In this paper, we aim to examine the effectiveness of a social gamification mechanism, named user scores, designed in Foursquare which is one of most popular location-based social networks. A user’s score in Foursquare is an aggregate measure based on recent check-in activities of the user, which reflects a snapshot summary of the user’s temporal and spatial behaviors. Whenever a user checks in to a venue, a list of scores of the user’s friends are visible to the user via a ”leaderboard” which ranks these users’ scores in a descending order. Given a pair of friends who participate in a score competition in such a gimification mechanism, we identify if one user’s scores have significant influence on the other user’s scores by utilizing the Granger Causality Test. To understand what types of users and what types of friends tend to participate in the score competition (i.e., their check-ins are more likely driven by such a gamification mechanism), we extract users’ features (e.g. user’s degree) as well as the features of pairs of friends (e.g., number of common friends, score similarity and ranking difference) to examine whether these features have correlations with those pairs of users who are identified as being involved in the score game. The identified influence on user scores has the important implication on applications including friend and venue recommendations in location-based social networks.