Advanced Materials Research Vols. 765-767

Abstract: Persona is a virtual describe of some similar users. It plays an important role in product development and web design, etc. Persona is always described in narrative form such as words and photos. This paper provides a new method of describing persona with digitization, which is to record and analyze users behavior features in real time through the interaction of browser and server, and calculate the user's persona accordingly. So it will give a rapid and exact estimate in user identification and user clustering, etc.

Abstract: String matching algorithm is one of the key technologies in numerous network security applications and systems. Nowadays, the increasing network bandwidth and pattern set size both call for high speed string matching algorithm for large-scale pattern set. An improved algorithm based on WM algorithm for large-scale pattern set is proposed in this paper. The presented multiple pattern string matching algorithm, DWMH, which we call in brief, combines the idea of Horspool algorithm with WM algorithm and applies the method of double hash to revise WMs HASH table to achieve better performance. Our extensive experiments demonstrated that DWMH algorithm is more efficient than WM algorithm, particularly when the size of pattern set becomes large-scale.

Abstract: In this paper, we integrate the DMA data transmission technology and memory mapping technology to implement zero-copy in the Linux operating system. Zero-copy allocates memory address space to network card. In this case, the network card directly transmits the received data to the memory space allocated in advance when the network card transmits the data to the Linux kernel by using DMA. After that, zero-copy maps the allocated memory space to the process space of the user application. And then, the user application can access this allocated memory space directly and reduce the memory copies between the kernel and user space.

Abstract: Could computing is a new type of computing mode, it distributes computation task on the resource pool which consists of massive computers, so the application systems can obtain the computation power, the storage space and software service according to its demand. The understanding of cloud computing is developing and changing constantly, cloud computing still has no unanimous definition. This paper describes three main service forms of cloud computing: SAAS, PAAS, IAAS, compared the definition of cloud computing which is given by Google, Amazon, IBM and other companies, summarized the basic characteristics of cloud computing, and emphasized on the key technologies such as data storage, data management, virtualization and programming model.

Abstract: The rapid development of internet technology leads to an effective way to share ideas with digital information. Extracting text information from digital information especially complex background images has become increasingly important due to the demand of text understanding. In this paper, we propose a k-means clustering based method for text extraction. After locating text regions with the corner detection, an improved k-means algorithm is adopted to extract text from complex background. Experiment results show the high performance of the presented method.

Abstract: This article focused on the energy limit property of Wireless Sensor Network, and proposed a residual energy based algorithm WN-LEACH, with the classic network mode of LEACH routing algorithm. The algorithm combines the proportion of residual energy in the total energy with the cumulative number of the normal nodes supported by the cluster heads as a cluster selection reference. In order to balance the energy consumption of each cluster-head, the algorithm took both the different positions of the base station and the initial energy of the network into consideration, and weighted the two factors to balance the energy consumption between transmitting the signals and data fusion. Simulation results show that the algorithm can promote the lifetime of the uneven energy network and does not impair the effects of the LEACH algorithm.

Abstract: Ant colony optimization algorithm is a very important one among swarm intelligence algorithms. Because Java is an advanced object-oriented and platform-independent computer programming language, in order to use this algorithm in a platform-independent and flexible way, this paper introduces a Java-based implementation package of it. This package includes some sub-packages. There are several classes which are all implemented in Java using object-oriented technology in each sub-package. Users can utilize these classes on computers installed the corresponding Java runtime environment to solve some problems. After the test on two travelling salesman problems, these classes performed properly and efficiently, and the good effect was received.

Abstract: This paper presents a novel collaborative filtering recommendation algorithm based on field authorities which simulates the real life word of mouth recommendation mode. It uses the specialistic knowledge from field authorities of different genres, and successfully addresses data sparsity and noise problems existing in traditional collaborative filtering. Meanwhile it also improves prediction accuracy and saves computational overhead effectively. Experiments on MovieLens datasets show that the accuracy of our algorithm is significantly higher than collaborative filtering approach based on experts, and has larger scope because of no external data limitations. Meanwhile, compared to traditional k-NN collaborative filtering, our algorithm has a better performance both in MAE and precision experiments, and the computational overhead has a decrease of 19.2% while they provide the same accuracy level.

Abstract: With the explosive growth of mobile malware in mobile internet, many polymorphic and metamorphic mobile malware appears and causes difficulty of detection. A mobile malware network behavior data mining method based on behavior categorization is proposed to detect the behavior of new or metamorphic mobile malware. The network behavior is divided into different categories after analyzing the behavior character of mobile malware and those different behavior data of known malware and normal action are used to train the Naïve Bayesian classifier respectively. Those Naïve Bayesian classifiers are used to detect the mobile malware network behavior. The experiment result shows that Behavior Categorization based Naïve Bayesian Classifier (BCNBC) can improve the detection accuracy and it can meet the requirement of real time process in mobile internet.

Abstract: In view of the present personalized ranking of search results user interest model construction difficult, relevant calculation imprecise problems, proposes a combination of user interest model and collaborative recommendation algorithm for personalized ranking method. The method from the user search history, including the submit query, click the relevant webpage information to train users interest model, then using collaborative recommendation algorithm to obtain with common interests and neighbor users, on the basis of these neighbors on the webpage and webpage recommendation level associated with the users to sort the search results. Experimental results show that: the algorithm the average minimum precision than general sorting algorithm was increased by about 0.1, with an increase in the number of neighbors of the user, minimum accuracy increased. Compared with other ranking algorithms, using collaborative recommendation algorithm is helpful for improving webpage with the user interest relevance precision, thereby improving the sorting efficiency, help to improve the search experience of the user.