Vol. 35 · No. 5

Traditional recommender systems provide users with customized recommendations of products or services. They employ various technologies and algorithms in order to search and select the best options available while taking into account the user's context. Increasingly often, such systems run on devices in heterogeneous environments (including mobile devices) making use of their functionalities: various sensors (e.g. movement, light), wireless data transmission technologies and positioning systems (e.g. GPS) among others. In this paper, we propose an innovative recommender system that determines the best service (including photo and movie conversion) and simultaneously accommodates the context of the device in a heterogeneous environment. The system allows the choice between various service providers that make their resources available using cloud computing as well as having the services performed locally. In order to determine the best possible recommendation for users, we employ the concept of learning agents, which has not been thoroughly researched in connection with recommender systems so far.

In the constantly growing blogosphere with no restrictions on form or topic, a number of writing styles and genres have emerged. Recognition and classification of these styles has become significant for information processing with an aim to improve blog search or sentiment mining. One of the main issues in this field is detection of informative and affective articles. However, such differentiation does not suffice today. In this paper we extend the differentiation and suggest a fine-grained set of subcategories for affective articles. We propose and evaluate a classification method employing novel lexical, morphological, lightweight syntactic and structural features of written text. The results show that our method outperforms the existing approaches.

Multi-Agent theory which is used for communication and collaboration among focused crawlers has been proved that it can improve the precision of returned result significantly. In this paper, we proposed a new organizational structure of multi-agent for focused crawlers, in which the agents were divided into three categories, namely F-Agent (Facilitator-Agent), As-Agent (Assistance-Agent) and C-Agent (Crawler-Agent). They worked on their own responsibilities and cooperated mutually to complete a common task of web crawling. In our proposed architecture of focused crawlers based on multi-agent system, we emphasized discussing the collaborative process among multiple agents. To control the cooperation among agents, we proposed a negotiation protocol based on the contract net protocol and achieved the collaboration model of focused crawlers based on multi-agent by JADE. At last, the comparative experiment results showed that our focused crawlers had higher precision and efficiency than other crawlers using the algorithms with breadth-first, best-first, etc.

Using MANETs for real time applications is always a challenge as the network is extremely dynamic with brisk topology changes. Despite this, several real time schedulers have been developed that aimed at providing QoS to ad hoc nodes. The quality of service (QoS) is standardized in terms of capacity, reliability, link quality, delays/jitters, and network cost. Thus, for QoS, the better transmission should be maintained at end user as well as at the transmitting unit. QoS of a network is affected by delays and bandwidth allocated for transmission. For an efficient network, it is required to predict these metrics during transmission. For this, in this paper, integration of quaternion-based Kalman filter is performed that predicts the required bandwidth and the network delays with higher accuracy. From the analysis, it is shown that bandwidth can be optimized but it is not possible to aloof delays in the network. Thus, while implementing such admission control procedures, estimation process allows control over delays and sustain them from going beyond a certain threshold value. The model proposed is a novel approach and has not been formulated in any of previous work related to QoS in MANETs. The effectiveness of model is demonstrated using both simulation and real time results.

The aim of this paper is to build an online failure prediction classifier for monitoring the behavior of programs. The classifier predicts the termination state of the program execution paths as failing or passing. This could be achieved by mapping each execution path as a vector into a feature space whose dimensions represent common sub-paths amongst failing and passing execution paths. The main contribution of this paper is to treat the failure prediction problem as a classification task of execution paths in a customized feature space. The main dilemma is the size and the number of space dimensions, affecting the speed of the classifier. The size of the dimensions could be reduced by shortening the length of the common sub-paths, used as the space dimensions. The length of common sub-paths is affected by repeated patterns in program executions. Replacing the consecutively repeated patterns with only a single iteration in execution paths, reduces the size of the common sub-paths. The number of dimensions could be reduced by removing dimensions which have projection onto others. This paper proposes two kernels which measure similarity amongst execution paths in an implicit feature space with reduced dimensionality. Our experiments demonstrate a significant reduction in time overhead of the failure prediction classifier while preserving accuracy.

Automated and intelligent video surveillance systems play an important role in the modern world. Since the number of various video streams that must be analyzed concurrently grows, such systems can assist humans in performing tiresome tasks. In order to be effective, video surveillance systems have to meet several requirements: they must be accurate and able to process the received video stream in real-time. A robust system should not depend on lighting conditions, illumination changes and other sources of scene variation. A common component of surveillance systems is a module that performs background estimation and foreground segmentation. The MoG (Mixture of Gaussians) algorithm is a widely used statistical technique of video segmentation. The estimation process is time-consuming, especially for complex mixture models containing many components. The work presented here focuses on the performance evaluation of MoG algorithm aiming to assess feasibility of OpenCL-based processing of high resolution video on GPU accelerated platforms.

Privacy preserving data publication is a major concern for both the owners of data and the data publishers. Principles like k-anonymity, l-diversity were proposed to reduce privacy violations. On the other side, no studies were found on verification on the anonymized data in terms of adversarial breach and anonymity levels. However, the anonymized data is still prone to attacks due to the presence of dependencies among quasi-identifiers and sensitive attributes. This paper presents a novel framework to detect the existence of those dependencies and a solution to reduce them. The advantages of our approach are i) privacy violations can be detected, ii) the extent of privacy risk can be measured and iii) re-anonymization can be done on vulnerable blocks of data. The work is further extended to show how the adversarial breach knowledge eventually increased when new tuples are added and an on the fly solution to reduce it is discussed. Experimental results are reported and analyzed.

The astonishing rapid diffusion of portable devices (i.e. smartphones, tablets, etc.) has had a big, and often positive, impact on our every-day life. These devices have new advanced features developed specifically because of user demand. For example, it is now possible to publish directly the pictures obtained by means of the internal camera of a smartphone on our social network accounts, or on an image hosting service. It is therefore important to have tools, on the portable devices, that can prove the ownership of the pictures and to use them before publishing images. Digital watermarking techniques are commonly used for the copyright protection of images and videos. We have developed a tool for portable devices based on the Android OS that allows the embedding of a digital visible or invisible watermark into a digital image.

Certificate-based public key cryptography not only solves certificate revocation problem in traditional PKI but also overcomes key escrow problem inherent in identity-based cryptosystems. This new primitive has become an attractive cryptographic paradigm. In this paper, we propose the notion and the security model of certificate-based designated verifier signatures (CBDVS). We provide the first construction of CBDVS and prove that our scheme is existentially unforgeable against adaptive chosen message attacks in the random oracle model. Our scheme only needs two pairing operations, and the signature is only one element in the bilinear group G 1. To the best of our knowledge, our scheme enjoys shortest signature length with less operation cost.

Nonparametric Discriminant Analysis (NDA) possesses inherent advantages over Linear Discriminant Analysis (LDA) such as capturing the boundary structure of samples and avoiding matrix inversion. In this paper, we present a novel method for constructing an updated Nonparametric Discriminant Analysis (NDA) model for face recognition. The proposed method is applicable to scenarios where bursts of data samples are added to the existing model in random chunks. Also, the samples which degrade the performance of the model need to be removed. For both of these problems, we propose incremental NDA (INDA) and decremental NDA (DNDA) respectively. Experimental results on four publicly available datasets viz. AR, PIE, ORL and Yale show the efficacy of the proposed method. Also, the proposed method requires less computation time in comparison to batch NDA which makes it suitable for real time applications.