Earticle

The staggering growth in Internet of Things (IoT) technologies is the key driver for generation of massive raw data streams in big data environments. This huge collection of raw data streams in big data systems increases computational complexity and resource consumption in cloud-enabled data mining systems. In this paper, we are introducing the concept of pattern-based data sharing in big data environments. The proposed methodology enables local data processing near the data sources and transforms the raw data streams into actionable knowledge patterns. These knowledge patterns have dual utility of availability of local knowledge patterns for immediate actions as well as for participatory data sharing in big data environments. The proposed concept has the wide potential to be applied in numerous application areas.

Using data mining technology to analyze the huge amounts of meteorological data plays an important role in improving the accuracy of weather forecasts. After analyzed the features of meteorological data, a distributed meteorological data mining models using the pattern structure in formal concept analysis is proposed in this paper. Since there exists large numerical, boolean, and geographic concepts in meteorological data, using classic methods of formal concept analysis needs to build single-valued formal context. This paper adopts concept lattice pattern structure to avoid such conversions and the results of rules mining have higher readability and efficiency. This pattern structure of concept lattice is extended to the distributed model to improve data processing capability.

The advent of social computing brought with it different social networking platforms. The idea of surfers socializing with people of different backgrounds as well as geographical regions is quite fascinating. In our approach, we delved deeper in disaster discovery whereby we extracted panic related attributes and trained them with real data in three disaster scenarios in different parts of the world. Fine tuning of the final attributes led to accuracies above 91% proving the fact that with proper attribute selection and handling of sparse data balance, it’s possible to detect related disasters as soon as related tweets appear. We believe that we are the first to use probabilistic classifiers approach as well as NLP in specifically human induced terror attacks detection as there is no known system currently that solely caters for these.

Twig query is a core operation in processing and optimizing XML structural queries. Recently, various algorithms have been proposed for finding twig patterns efficiently. Most of them are based on region labeling, pay little attention on nodes level information, and require additional caches such as stacks or lists to maintain the intermediate matching results, which cause the performance bottleneck of these algorithms. In contrast to previous work, we present a bottom-up algorithm for twig queries, which does not require additional caches and introduces idea of string searching to determine binary relationship between two nodes. Be- sides, this paper presents a node filtering mechanism–PathLevel, which can also be used in other algorithms for speeding up the query. Comprehensive experiments on several datasets demonstrate our method is an effective way for twig query.

Document similarity is a common task to a variety of problems such as clustering, unsupervised learning and text retrieval. It has been seen that document with the very similar content provides little or no new information to the user. This work tackles this problem focusing on detecting near duplicates documents in large corpora. In this paper, we are presenting a new method to compute pairwise document similarity in a corpus which will reduce the time execution and save space execution resources. Our method group shingles of all documents of a corpus in a relation, with an advantage of efficiently manage up to millions of records and ease counting and aggregating. Three algorithms are introduced to reduce the candidates shingles to be compared: one creates the relation of shingles to be considered, the second one creates the set of triples and the third one gives the similarity of documents by efficiently counting the shared shingles between documents. The experiment results show that our method reduces the number of candidates pairs to be compared from which reduce also the execution time and space compared with existing algorithms which consider the computation of all pairs candidates.

Discovery of loose data linkages between data services can help on-demand Web data integration in accordance with situation changes. However, we met the uncertainty challenge when discovering such data linkages with current automatic matchers. To handle the uncertainty problem, this paper develops a synthesized matching algorithm to combine the matching results from multiple automatic matchers with user feedbacks. It also proposes a service hyperlink model to encapsulate such data linkages for further reuse. Experiments show our approach can effectively improve the correctness of discovered data linkages.

Bacterial Foraging Optimization (BFO) is a novel optimization algorithm based on the social foraging behavior of E. coli bacteria, but it is difficult to optimize to get a high precision due to the randomness of the bacterial behavior, which belongs to intelligence algorithm. This paper presents an extended BFO algorithm, namely the Cooperative Bacterial Foraging Optimization (CBFO), which significantly improves the original BFO in solving clustering problems. A novel clustering method based on the CBFO could be used for solving clustering problems. In this work, firstly, The efficiency and performance of the CBFO algorithm was evaluated using six widely-used benchmark functions, coming up with comparative results produced by BFO, then Particle Swarm Optimization (PSO) is studied. Secondly, the algorithm with CBFO algorithms is used for data clustering on several benchmark data sets. The performance of the algorithm based on CBFO is compared with BFO algorithms on clustering problem. The simulation results show that the proposed CBFO outperforms the other three algorithms in terms of accuracy, robustness and convergence speed.

In this paper, the use of finite Gaussian mixture modal (GMM) based Expectation Maximization (EM) estimated algorithm for score level data fusion is proposed. Automated biometric systems for human identification measure a “signature” of the human body, compare the resulting characteristic to a database, and render an application dependent decision. These biometric systems for personal authentication and identification are based upon physiological or behavioral features which are typically distinctive, Multi-biometric systems, which consolidate information from multiple biometric sources, are gaining popularity because they are able to overcome limitations such as non-universality, noisy sensor data, large intra-user variations and susceptibility to spoof attacks that are commonly encountered in mono modal biometric systems. Simulation show that finite mixture modal (GMM) is quite effective in modelling the genuine and impostor score densities, fusion based the resulting density estimates achieves a significant performance on eNTERFACE 2005 multi-biometric database based on dynamic face and signature modalities.

With the promoting of big data analysis in every industry, the medical industry is also being widely used big data analysis as the tools, medical industry accumulated a large amount of data because of the wide ranging, it’s has an important role in creating new business value for the medical industry and enhancing healthcare industry by performing the analysis of big data,. This paper focuses on the analysis of big data applications in the medical industry, and discussed the potential of its commercial value for the healthcare industry.

Most time series big data is with noise and uncertain. To abstract the key information effectively and quickly, the estimation is one of the feasible methods for the uncertain big data. The Kalman filter with adaptive method by part of samples can give the high dimensional characteristics, reduce the computing cost and data uncertainty, but encounter the irregular estimation. The number of sample and the performance of the abstracted information have the tradeoff, which means we can use the suitable number of sample to abstract the key information of the series data. This paper discusses how to find the suitable sampling points for the time series data and the simulations show that the key dynamic information of time series big data can be guaranteed with the compression amount number of sample data.

As technology advances, mobile phones have changed significantly with devices and operating systems becoming more sophisticated. Mobile Applications (Apps) have been increasingly popular in the recent years and are changing the people’s daily lives in leisure and businesses. The Booming industry of Apps makes great profit and the awareness to prevent customers or users from feeling dissatisfied is, accordingly, an important issue. Customer dissatisfaction may cause switching behavior, decreasing loyalty, and negative word-of-mouth among customers which may be the potential problem of business losses. Scarce researches have been done in discussing the effects of dissatisfaction among mobile products especially in applications issues. The purpose of this research is to propose a conceptual model and make an empirical investigation about the elements influencing customer dissatisfaction on using mobile applications. The sample consists of 200 respondents by using online questionnaires to collect data. The analysis employing structural equation modeling (SEM) shows that functionality, perceived usefulness and content have significant impacts on customer dissatisfaction. Implications for managerial perception and future research are discussed.

The knowledge transfer learning can generalize across domains where the types of objects and variables are different. Previous studies ignored connectivity and creativity of domain knowledge. Thus, these studies just transfer knowledge from a source domain to a target domain that not effectively use the knowledge from other domains. We proposed a method, called Multi-domain second order knowledge integration (MSKI), for integrating to address this problem. We hybridize and create new knowledge, which is formalized into an uncertain hypergraph. Then, we proposed a method to mine frequent sub-hypergraph from the uncertain hypergraph (MFS-UHG). The frequent sub-hypergraphs are pivot knowledge, which has to be transferred with high priority. We embed the pivot knowledge in the progress of MLN structure learning. The experimental evaluation on four domain datasets shows that the MSKI outperforms state-of-the-art MLN-based transfer learning.

The time aspect is not currently taken into account for finding a region of interesting (ROI) or a hot region, so that due to the time to visit frequently a place cannot be determined, it is difficult to discover the visiting regularity for a moving object. To this end, the spatio-temporal item (STI) and frequent spatio-temporal item (FSTI) integrated spatial and temporal attributes are defined. The FSTIs can represent a moving object often visits which area in what time, which can provide more useful information to improve the level of the location-based services(LBS). In order to find FSTIs, STIs are generated by using a density-based clustering algorithm to recognize the stay regions of objects, and then the STIs are mapped to 3D-grids integrated spatial and temporal dimensions. Finally, the extraction - merger strategy is used on the frequent grid cells to recombine the FSTIs. Experimental results on real dataset show that the approach proposed for mining FSTIs is effective.

Ontology has become a hot research topic in the fields of artificial intelligence such as knowledge representation, knowledge engineering and natural language processing (NLP) etc.. In this paper, according to the application requirements in the intelligent Uyghur information retrieval system, by giving the brief description about the ontology and its construction rules, methods, tools and descriptive languages, have conducted the contrastive analysis the current research status about the ontology in domestic and abroad, and then sum up some key issues in Uyghur ontology construction procedures and some early achievements. After all, the further research directions are also proposed in this paper.

XML has become the de facto standard for data exchange and transformation on the World Wide Web and is widely used in many applications of various fields, so it is urgent to develop some efficient methods to manage, store, query, and update XML data. There are two main methods to do this: the first method is a native approach which uses native XML databases to store XML data, and the second method use other mature commercial databases approaches to store and manage XML data considering the advantages of mature technologies of the commercial databases, especially use relational databases to store, query, and update XML data. For relational databases approach, although it can take advantage of mature technologies of relational databases, it needs to map XML data to relational data. In this paper, we research the problem of how to store XML data so that storing and updating of original XML data can be efficient than relational approach. We proposed a method to store XML data into linked lists with inverted index, in which the relationships between nodes of XML data tree are preserved by the links in linked lists. Inverted index are created for linked lists for efficiently querying and updating XML data tree. Two kinds of updates are considered including inserting a new node in or deleted an existed old node from XML data tree. Theoretical analysis of our algorithms shows that the methods proposed in the paper are efficient.

This paper studied the techniques of distributed data storage in wireless senor networks. Firstly, the challenge and the need for such techniques were summarized; Secondly, some representative distributed data storage and retrieval schemes were introduced in detail; finally, the future research directions and open issues were pointed out.

In recent years, cloud computing is highly embraced and more organizations consider at least some type of cloud strategy and apply theming their business process. Since failure is probable in cloud data centers and access to cloud resources available is fundamental, evaluation and application of different fault-tolerance methods is inevitable. On the other hand, the increasing growth of cloud storage users motivated us to study fault-tolerance techniques, and their strengths and weaknesses. In this paper, after introducing the concept off ault-tolerance in the context of cloud computing, the fault-tolerant techniques are presented, and after introduction of some measures, a comparative analysis is provided.

Owing to the complex structure and multi-meaning, property alignment is generally regarded as a challenging problem in the context of linked data. In this paper, we propose a novel method to align properties between datasets of linked data. Considering the role of properties in RDF triples, we regard all properties of linked data as property functions, and convert the problem of property alignment to the similarity evaluation between property functions, while the equivalent instances as inputs of property functions. Based on the similarity of property functions, the property alignment process of linked data is introduced. In order to prove the validity, we use the method to align properties in five representative domains between DBpeida and YAGO, DBpedia and LinkedGeoData respectively. The experimental results show that our method is independent of the property naming rules and can retrieve some matching properties ignored by other methods. In addition, our method requires fewer entity co-reference links than the link statistical approach.

Current trends in software system, such as cloud and big data platform, are leading to rapid and continuing changes. At the same time, these systems will have to react to these changes at runtime to satisfy the potential Quality of Service (QoS). Self-adaptation is recognized as a practical way for a software system to meet QoS requirements. The Development of self-adaptive software is generally more challenging and more difficult due to their high complexity. To address these challenges, this paper reviews the related research of self-adaptive software system and reports a case study that investigates a self-adaptive concurrency controller for database system. Through the case we illustrate how to develop a self-adaptive software system. Compared with other traditional method, the experimental results demonstrate that our self-adaptive controller can effectively improve the database performance by adjusting the MPL value based on workload changes and QoS requirements. Finally some future trends in this area are prospected and discussed.

With wide spread use of digital data communication on networks, security of data has become an important issue which provides integrity and confidentiality of content that ensures the distribution of information appropriately. While dealing with encoded content in the context of XML, its hierarchical tree structure has different level of confidentiality and integrity for various portion of the same content. Thus it imposes the need of dissemination approach specifically tailored for XML data to address the issues of efficiency and scalability. However these characteristics must be achieved without compromising the security and privacy of contents. The goal of efficient proposed Signature scheme for secure & selective distribution of XML content is to provide uniform platform for different data representation for secure and efficient availability of data. Proposed Signature Scheme is used to detect the change of content at information providers to discover and deliver new content to users. This can be achieved using structural properties of hierarchical tree structure data model with multicast topology approach.

3D motion capture is to track and record human movements. In recent years, it has been applied into many fields, such as human computer interaction, animation, etc. Low-level segmentation of motion capture data is of significance to the various applications of 3D motion capture; however, due to the high dimensionality of motion capture data, traditional low-level segmentation methods can hardly work out a suitable segmentation for motion capture data. In order to solve this problem, a low-level temporal segmentation algorithm based on cosine distance is proposed, hierarchical clustering is explored so that similar velocity vectors are clustered together according to the cosine distance in a progressive way, the center of each cluster is updated as the vector derived with linear regression, the segment boundaries are determined as the point when the cosine distance between adjacent velocity vectors is greater than 1 (angle>90 degrees). We have conducted experiments on the motion capture database provided by Carnegie Mellon University (CMU), the experiment results show that the performance of the proposed method is optimistic.

Decision problem of intersection checking for regular expressions plays an important role in the XML type checking. The typical technique is converted into the problem of automata intersection, which may generate a lot of redundant computing during the conversion. In the present paper, according to the features of XML schema languages, a new intersection checking algorithm based on inference system for regular expressions is proposed. This method is derived directly based on regular expression without the need for any conversion. For general regular expressions that is exponential time algorithm, but without constructing automata and for some special cases, especially for the one-unambiguous regular expressions used in XML type checking, is the polynomial time algorithm. Proofs of the correctness and completeness of the inference rules are given. Experiment results show that our approach are more effective than automatic approach in practical.

Cloud computing provides virtualization of services, has enabled opportunities for integration and collaboration of many e-businesses around world. Also, semantic web technology has been evolving in parallel, which uses RDF and OWL as storage formats for business data. Integration of semantic web technology on cloud platform provides an efficient way of managing data for e-commerce based applications. Many databases storage approaches have been proposed for e-commerce applications. However, there is a need to model the data storage based on semantic information so that the access to data store based on semantic behavior of user context and past history. The proposed work models a mobile phone e-commerce application which uses RDF data store which is maintained in cloud environment. SPARQL queries are used for access and retrieval mechanism.

Data warehouse is subject-oriented organized. However, when data warehousing, the hierarchy structure of subject is currently only decided by decision makers’ intuition. Faced with complicated business data mining cases, it is hard to establish hierarchy structure of subject just according to intuition. Thus a method based on ISM is present to make subject level structure establishment more measurable and illustrative. In this article, the "Subject" level structure establishment process is presented firstly. Then the method is put forward. Finally, the rationality and validity of this method are verified by a case on university financial data warehouse’s subject level establishment.

Sentence-level aligning bilingual parallel corpus is shown significant and indispensable status in machine translation, translation knowledge acquiring and bilingual lexicography research fields, which is the fundamental work for natural language processing. Given the great deal of work in sentence alignment and a variety of methods have developed for bilingual terminology extraction, those are unpractical for newly underway Tibetan information processing because those methods have to use a large number of manufactured sentences as training corpus while extracting inter-translatable word pairs. This paper proposes a multi-strategy Tibetan-Chinese sentence alignment method based on length of sentence, syntactic rules and bilingual dictionary. We test our approach on a bilingual corpus crawled from bilingual website and perform manual evaluation on bilingual sentences pairs extracted from Tibetan-Chinese corpora.

Support vector machine is a machine learning method. It takes on the good generalization ability and prediction accuracy. But the parameters of SVM model seriously affect the generalization ability and prediction accuracy of SVM model on the great extent. So an improved particle swarm optimization (PSO) algorithm based on chaotic search is introduced into the SVM model in propose a novel data classification (AMPSVM) method for processing the complex data. The first, the ergodicity, stochastic property, and regularity of chaos is used to chaotically search the current best individual, which randomly replaces the selected individual in the population in order to speed up evolution, improve the searching ability, convergence speed and accuracy. Then the improved PSO algorithm is used to select and optimize the parameters of the SVM (AMPSVM) model in order to improve the learning performance and generalization ability of the SVM model. In order to verify the effectiveness of the AMPSVM method, UCI data is selected in here. The experiment results show that the proposed AMPSVM method takes on the strong generalization ability, best sensitivity and higher classification accuracy.

Data mining extracts knowledge or patterns from a large amount of data. Secure communication is an issue of shared database applications. Fundamentally, secure multi party computation is to enable a number of networked parties to carry out distributed computing tasks on sensitive information. In this paper, we have proposed two techniques for multi party communication one is for third party assisted and another without third party. Third party assisted technique uses Group based approach and in case of without third party mechanism ECC based approach is used. Simulation results show that the time taken by the schemes are significantly less as compared to other schemes and this proves the efficacy of the proposed scheme which makes it viable for multi party communication in data mining applications.