Applied Mechanics and Materials Vols. 336-338

Abstract: Recent research on knowledge representation demonstrated that the more clearly domain concepts, the higher accuracy of the information extraction. Accounting that the ontology with an advantage in the aspects of the concept representation, it can be introduced to the Salt Lake domain to construct the domain ontology, and will convenience for the future of information extraction and retrieval. The main contribution of this paper is to propose a process-oriented domain ontology construction method which through comparative analysis of the domestic and international typical domain ontology construction method, and combined with the characteristics of the Salt Lake field, not only more comprehensive representation of the characteristics of the domain, but also has better logic and relevance.

Abstract: Ontologylearning has become a research hotspot in computer field at present with focuson concept and relation acquisition. The acquisition results of traditionalstatistic-based method are not highly accurate because it ignores the effectsof conceptual semantic and characteristics and others factors on conceptualrelation. To resolve above problems, this paper puts forward to a comprehensiveontology similarity integrates semantic similarity with characteristic attributesimilarity, takes uses of association rule to mine concept-to-conceptrelations, and applies comprehensive ontology similarity into relationacquisition. The experiment shows that this method can effectively improve theaccuracy of concept relation acquisition.

Abstract: Data fault tolerance is a key technology in the field of distributed storage. In this paper, an algorithm to encode massive amounts of data and then distribute storage these data on each node in the data center is proposed, aiming at coping with the serious challenges in the protection of data fault tolerance. The method converts multiplication operation in Cauchy RS coding into a binary multiplication through the transition on bit operation, so that the entire operation on RS encoding is converted to an operation containing only simple XOR operator. The experiment proves that the method is better than the copy and the original RS coding in the data encoding efficiency. Furthermore, it saves the storage space and promotes the application of erasure codes strongly in distributed storage field.

Abstract: Faced up with the fact that the buoy used in marine environmental monitoring did not install the self-destruction mechanism, and combined with the current development of the data destruction technology, data self-destruction technology for buoy was researched. The study shows that when the buoy system is out of control, it can select the two different data self-destruction methods by analyze different degree of threat which buoy system suffered. So it can destruct the marine environmental factor data in a fast and effective way, and make the data be unrecovered, thereby prevent data leaks, ensure data security.

Abstract: Krista Rizman Zalik and Borut Zalik proposed indices SV and OS associating with separation and compactness or overlap. Compactness and overlap were calculated by a few data points of a cluster, which makes the indices able to identify the number of clusters underlying the data set of different sizes and densities. However, the measure of overlap depends on two undefined variants and the measure of compactness are build on randomly selected ten percent data points of a cluster, which makes them difficult to compute and unpractical. This paper supposes to measure the compactness using ten percent of data points of a cluster that are farthest away from the center of the cluster, and revises the measure of overlap so that two undefined variants are eliminated from the measure of overlap. Experiments show that the modified index SV can identify the optional number of clusters underlying the data set of different size and prefers to the index OS.

Abstract: Data mining technologies are used widely while the side effects it incurred are concerned so seriously. Privacy preserving data mining is so important for data and knowledge security during data mining applications. Association rule extracted from data mining is one kind of the most popular knowledge. It is challenging to hide sensitive association rules extracted by data mining process and make less affection on non-sensitive rules and the original database. In this work, we focus on specific association rule automatic hiding. Novel strategies are proposed which are based on increasing the support of the left hand and decreasing the support of the right hand. Quality measurements for sensitive association rules hiding are presented.

Abstract: While crowdsourcing is developed as a business model, it has become a new method of information exchange under the development of mobile Internet and SNS, and promoted the method of information production and exchanging on Internet. During the processes of global huge disaster relief, mass of information released by normal people have been used and played an important role. Using crowdsourcing in disaster information collection can improve the efficiency of rescue. Based on the analysis of current status of disaster information collection and crowdsourcing model, this paper studied the crowdsourcing-compatible disaster information management system, including its content, architecture and future development direction.

Abstract: The problem of capitalizing on experience of how to track and analyze the learning processes is of particular interest in Computer Supported Collaborative Learning (CSCL) environments, where learning is usually based on the interactions occurring among peers. This paper provides some examples of DPs addressing tracking problems in CSCL environments. Our DPs have been elaborated with a twofold aim to contribute to the field by proposing possible solutions to tracking problems frequently faced in CSCL; on the other hand, by using DPs as a tool for sharing experience, we hope to inform the development of functions that could be integrated in Computer Mediated Communication (CMC) systems in order to support the monitoring, evaluation and assessment of the learning experience and therefore augment its effectiveness.

Abstract: This paper proposes the classification algorithm of news pages based on domain Ontology. In order to improve the shortage of current classification algorithm that only considers the content similarity, this paper presents the semantic classification method which considers both content similarity and structural correlation. Firstly, it parses the Ontology to get Ontology category vector, extracts keywords of news pages texts and drops semantic dimension. At this time, finding out the same vocabulary and ontology category vector in page texts to constitute the text expectation vector, and then calculating the content similarity between ontology category vector and expectation vector of text by using the law of cosines. Secondly, the common vocabularies are mapped to the ontology hierarchy chart, and the structural relevancy is obtained by calculating weighted path of this directed acyclic graph. Finally, it calculates the correlation degree of the news pages and Ontology by combining both, and determines the category of news pages by judging the size relationship between the result and the initial threshold value.

Abstract: The data in computer is lost because of the destruction of disk partition table. To solve this problem, a data recovery solution by recovering partition table was proposed. The hard disk logical parameter was introduced. The hard disk addressing mode was analyzed. On the basis of analyzing disk partition structure, the recovery principle of MBR and EBR was proposed. After the specific process of partition table recovery was illustrated, an implement scheme was proposed and proved to be available.