Applied Mechanics and Materials Vols. 347-350

Abstract: Algorithmic relationships are discovered here for programming tutoring. There are two kinds of algorithmic relationships between programming resources on the web: associative relationship and structural similarity relationship. They can be organized as a hierarchical body. An algorithm can solve different programming problems and a programming problem also can be solved by different algorithms. Thus, there is such algorithmic relationship, or associative relationship, between these programming resources on the web. The algorithmic structures of source codes can be mined by neural computing. Different source codes may have a structural similarity relationship between them, meaning that they are similar in their algorithmic structures. A learner can learn algorithms from simple to complicated structures or from similarities in their structures. In our experiment, we use a tree structure to organize the algorithmic relationships.

Abstract: Multi-dimensional applications use tree structure to store data and space filling curves to traverse data. Most frequently used Quad-tree and Z-ordering curve are analyzed. By importing these to a HDF5 file format, a multi-dimensional data storage subsystem is constructed. Performance test results show in a sequential reading application environment, this method is feasible and efficient.

Abstract: In this paper, we use GA to improve the D-S evidence theory, and apply the improved D-S evidence theory to VIP intelligent recognition and recommendation system. In the VIP intelligent recognition and recommendation system of clothes, there are three main evidences: body size, personal preferences, and purchase records. So collision often happens inevitable. This requirement asks us to find out a suitable method to identify the VIPs needs. D-S evidence theory can improve the rate of identification, but has no idea about the collision. The improved D-S evidence theory based on genetic algorithm can deal with the collision evidence and improve the rate of the identification and the stability. As such we can provide VIP more suitable recommendation. The experiment results of clothes recommendation demonstrate the flexibility of the improved method.

Abstract: In order to set up universal and non-linear map of variables, a full binary tree is constructed as mathematical model. Leaf nodes of the full binary tree are linear combination of input variables, and used as inputs of next nodes. On the basis of weighting two inputs by selector for inner node, the inputs are again linearly combined and used as output for next node. The inputs and outputs of all the inner nodes are constructed in turn as the same, and the output of root node is the output of mathematical model, implementing segment-linear approximation. With the means of machine learning of particle swarm optimization for data from some areas, all the coefficients of mathematical model are achieved for the special. The mathematical model is applied to seismic inversion to interpret stratum by seismic data, approving it very practical.

Abstract: It is always a challenge by using statistical method in corpus database to analyze semantics of natural language Sentences (NLS). This paper proposes a method of recognizing and translating ontology query in natural language, called OntoQuery-NLP. With the help of pre-create semantic templates, the OntoQuery-NLP maps NLSs matching the format of the semantic templates into formal semantic expressions. By parsing these semantic expressions, the OntoQuery-NLP recognizes the queries and gets the correct answers from ontology. Compared with other methods, the OntoQuery-NLP, without the support of any corpus, has faster retrieving speed and higher retrieving accuracy.

Abstract: t is one of important topics in social network research how to form member clustering according to potential social relations by automatic context-aware the user's behavior feature. This paper presents a context-aware mobile P2P social network framework, member clustering model and algorithm. The user's location information, environmental characteristics etc. are introduced to the clustering algorithm, which intelligently cluster to potential P2P social network. The experimental results show that the proposed approach and the algorithm have a higher response speed, load balance and adaptive ability.

Abstract: Trust management provides a potential solution for the security issues of distributed networks. However, there are rare researches about the trust mechanism for IoT in the literature. A new distributed trust management mechanism for IoT is established in this paper. Firstly, we extract three basic elements-service, decision-making and self-organizing, of trust management from the investigated trust solutions. Then, based on a service model, we establish a trust management frame-work for the layered IoT, which is decomposed into three layers: sensor layer, core layer and application layer. Finally, we use fuzzy set theory and formal semantics-based language to perform the layered trust mechanism. The proposed trust conception, layered service model and formal method provide a general framework for the study of trust management for the IoT, and further provide a significant reference for the development of sound trust models for IoT.

Abstract: In this paper, a resource scheduling algorithm for Long Term Evolution (LTE) downlink is proposed. It provides Quality of Service (QoS) guarantee to Guaranteed Bit Rate (GBR) services and also suit to the Non Guaranteed Bit Rate (Non-GBR). Different services possess different QoS needs in LTE system, based on this, 3GPP divide them into several types. In view of the division, we set priorities for different kinds of services to guarantee the transportation of GBR services which have high QoS needs. At the same time, we take the Media Access Control (MAC) layer logical channel priority into account, and consider users channel quality to acquire a higher cell throughput. Simulation results show that our scheme can improve the QoS of emergency services obviously at the cost of a certain amount of the non-real times.

Abstract: In this paper, a new precoding scheme is proposed based on the combination of Block Diagonalization (BD) and SLNR (Signal Leakage Noise Ratio) maximization. Then a new user selection algorithm is proposed based on the joint precoding scheme. BD precoding will cause performance loss in the single antenna terminals when the number of terminal antenna is inconsistent. The algorithm we proposed can overcome the drawback by using the maximum SLNR for single-antenna users and BD precoding for multi-antenna users respectively. Simulation results show that the proposed algorithm will enhance the system sum-rate performance significantly when SNR (Signal Noise Ratio) over 5dB. The performance improves by 30% when SNR reaches 20dB.

Abstract: The Web has become the largest information source, but the noise content is an inevitable part in any web pages. The noise content reduces the nicety of search engine and increases the load of server. Information extraction technology has been developed. Information extraction technology is mostly based on page segmentation. Through analyzed the existing method of page segmentation, an approach of web page information extraction is provided. The block node is identified by analyzing attributes of HTML tags. This algorithm is easy to implementation. Experiments prove its good performance.