Earticle

MapReduce is a popular framework for processing large datasets in parallel over a cluster. It has gained wide attention for its high scalability, reliability and low cost. However, its performance may be degraded by excessive network traffic when processing jobs, for such two problems as data locality in reduce task scheduling and partitioning skew. We propose a Minimum Transmission Cost Reduce task Scheduler (MTCRS) based on sampling evaluation to solve the two problems. The MTCRS takes the waiting time of each reduce task and the transmission cost set as indicators to decide appropriate launching locations for Reduce tasks. The transmission cost set is computed by a mathematical model, in which the parameters are the sizes and the locations of intermediate data partitions generated by Average Reservoir Sampling (ARS) algorithm. The experiments show that the MTCRS reduces network traffic by 8.4% compared with Fair scheduler.

A grouping of data objects such that the objects within a group are similar (or related) to one another and different from (or unrelated to) the objects in other groups. Many of clustering algorithm is available to analyze data. This paper intends to study and compare different clustering algorithms. These algorithms include K-Means, Farthest First, DBSCAN, CURE, Chameleon algorithm. All these algorithms are compared on the basis of their pros and cons, similarity measure, their working, functionality and time complexity.

Relational database can be replicated and distributed in the mobile computing scenario. To protect the ownership of the data and to trace the trail of data distribution, we devise a complex database watermark scheme. In each data distribution process, the sender and the receiver agree on a watermark key pair, and a watermark is embedded into the data. In case of piracy or misusing, the sender can claim the ownership of the controversial data copy and accuse the receiver for his improper behavior. On the other hand, an innocent receiver can prove his innocence to arbitral authorities with watermark detection. Experiment results show that the scheme has a satisfying performance and it is qualified for practical use.

The coalition formation is an important aspect of multi-agent negotiation and cooperation. Based on the Bilateral Shapley-Value, a multi-sided eigenvalue is presented by using the reference of n-person stochastic cooperative game. It is obvious when multi-sided eigenvalue have superadditivity, rational agents will combine to one coalition. A dynamic coalition formation algorithm is constructed based on the eigenvalue. Procedures of multi-sided negotiation, agent’s negotiation and coalition condensing are introduced in detail. In the end, the complexity, validity, coalition stability and parameter’s function of the algorithm is given. According to these, the correctness of the algorithm is proven.

Active database system has been introduced to extend the database functionality. It is capable of detecting specific events and automatically reacting to them by executing certain actions either inside or outside the database. This behavior is usually specified through Event Condition Action (ECA) rules. Rule design plays a critical role in implementing an active database system, it is not always straightforward due to lack of methodology. In this paper is presented a new approach to identify ECA rules that could strongly express application semantics. Additionally is proposed an ensemble method for triggering more appropriate ECA rules when the interesting events occurred.

Data management is the foundation of the University Human Resource Management Systems (UHRMSs). The data collected by the different departments in university human resource systems is distortion, logical confusion and unstructured. In order to overcome these defects, we design an adaptable university human resource management system. In the system, the technology of adaptive computing is adopted to enhance the self-management capability of UHRMS and decrease the complexity of management. Experimental results show that the proposed adaptable UHRMS has a lower transaction response time and a higher throughput than the traditional methods.

Nowadays with the rapid development of wireless sensor networks, and network traffic monitoring, stream data gradually becomes one of the most popular data models. Stream data is different from the traditional static data. Clustering analysis is an important technology for data mining, so that many researchers pay their attention to the clustering of stream data. In this paper, MSFS algorithm is proposed. By means of the experimental verification analysis, based on biologically inspired computational model, higher clustering purity on both the real dataset and the simulation datasets existence is demonstrated for the proposed algorithm. In other words, the cluster result of MSFS algorithm is advantageous over previous method.

With the explosive growth of the data stored in various forms, need for innovative and effective technologies to find and use information and knowledge from a large variety of data sources which is continually increasing. Web information contains a lot of noise. Web Mining is the application of data mining techniques to discover classification of web data. It focuses on techniques that could predict the data’s class while the user interacts with the web. The aim of this paper is to measure, propose and improve the use of advance web page classification techniques which is highly used in the advent of mining large web pages based data sets which allows data analysts to conduct more efficient execution of large scale web pages data searches. Thus in this paper researchers introduce an improved concept which may reduce the search space using classification techniques with optimization technique.

The data grid is one of the infrastructures for data and storage resource management. Data replication can improve data access performance, fault tolerance, and reduce network transmission bandwidth. As the site storage space is limited in the grid, how to effectively use grid resources become an important challenge. This paper introduces a dynamic grid replication algorithm based on popularity support and confidence (BPSC). Through the algorithm, data and its associated copy can be placed on a suitable site, with reduced access latency. Through Optorsim, simulation results show that the algorithm can provide better performance compared with other algorithms

In traditional data mining process, the definition of mining objects and analysis tasks are all decided artificially based on the analysts’ knowledge and experience. To achieve intelligent data analysis, a method called thinking theme discovery technology is proposed to imitate humans’ thinking models. Since traditional thinking theme discovery algorithm is based on hierarchical clustering, the efficiency of which is far from acceptable with the increasing of data amounts. This paper improves the efficiency of the algorithm on density-based clustering method. With five complex network datasets and one commercial theme dataset, the experimental results show that both the effectiveness and efficiency of the algorithm are improved.

Object categorization based on hierarchical context modeling has shown to be useful in large database of object categories, especially, when a large number of object classes needs to be recognized from a range of different scene categories. However, average precision of categorization is still low compared to other existing methods. This may reflect that the contribution of underlying relations between objects has not been fully considered. In this paper, we improve average precision of contextual object recognition by taking advantage of objects co-occurrence information. Our method consists of two main phases. In the first phase, object representation is derived by considering the frequency of objects appeared in each image. The second phase is focused on classification of objects by applying a decision tree algorithm. We use SUN09 database to evaluate our proposed method. This database consists of images spanning from different scene categories and object instances. Our experimental results demonstrate that our proposed method achieves a higher average precision in comparison to a recent similar method by encoding contextual information in an efficient way.

To solve the problem that the total number of files in HDFS is limited by the memory size of the NameNode, the paper proposed a scalable multiple NameNodes coexisting scheme for cloud storage. NameNode location service (NLS) was introduced into the system. NLS communicated with all NameNodes and was responsible for the file location requests from the clients. Experimental results showed this scheme had good scale-out scalability and the upper limit of the total number of files that the system was able to store could be greatly increased by adding a number of NameNodes.

Querying nested data has become one of the most pivotal issues for seeking desired information on the Web. Unlike the traditional information retrieval, to effectively manage nested data, we generally need store the data and its structures separately, which significantly reduces the performance of data retrieving, especially when the dataset is in a large scale. More seriously, it brings a big challenge on ensuring the efficiency of processing precise queries that need to locate the exact positions of some certain values in a nested dataset. Combining the techniques of column-strip storage and inverted index, this paper defines an expression to represent the data objects’ unique location in nested records— UPath, and based on which we present a new index structure— UniHash to support precise query processing on nested datasets. In addition, this work develops the related algorithms for building UPath, establishing UniHash, performing precise queries on UniHash with MapReduce platform and maintaining UniHash as well. Compared with some existing approaches, such as XPath-based and Dremel, UniHash index is capable of supporting the execution of precise queries over nested dataset with better performance. We give the results of a group of experiments, which were conducted on different real datasets, to demonstrate the efficiency of the approach.

It is proposed that this paper shows the types of pros and cons of storage and data retrieval tools. Also it shows the pros and cons of each type with a detailed explanation. The explanation shows how and when they can use any of them also determine the means and security applications in these tools and give the pros and cons of each one. In addition to provide tips and benefits of each one of them in terms of storage size and the number of tables. Also the number of users and analysis tools databases and support tools for each one of them, and how to repair, back up, and recover evidence rules for each one of them. Identifying the powers of users and the cons in each one and determine when Microsoft SQL server or Microsoft access can be used if needed. There are several versions of SQL server in the market. Each copy gives a certain capability. The user will find a copy for Project Enterprise. Which is the largest, the powerful, and most frequently used version. Also there is a developer’s copy. Where a developer will have an enterprise version with full specification. This version deigned only for developers who develop programs that run on SQL server. Moreover developer’s copy there is a miniature version of windows CE edition. Which works on windows CE systems for handheld devices.

The problems of the icons covering each other and slow data loading are very challenging when the GIS technology is applied to load a large number of topological nodes. A novel algorithm for aggregating the topological nodes in GIS based on the hierarchical clustering algorithm is presented in this paper. Comparison with the hierarchical clustering algorithm and partitioning algorithm is explored. In addition, the implementation of the algorithm is derived. The experiment demonstrates that the algorithm solved the problems of the icons covering each other and slow data loading effectively, realized the aggregation of the topological nodes preferably, and presented the distributed trend of the topological nodes and the topology relation between each other in the topology. The experimental results verify the effectiveness of the method.

In the present era as internet is growing with exponential pace, computer security has become a critical issue. In recent times data mining and machine learning have been researched extensively for intrusion detection with the aim of improving the accuracy of detection classifier. KDD CUP’ 99 Data set is the most widely used dataset in research domain. Selecting important feature on the basis of rough set based feature selection approach have lead to a simplification of the problem, faster and more accurate detection rates. In this paper, we presented an efficient approach for detecting relevant features from the KDD CUP’99 Data set.

During the last two decades, a substantial amount of research efforts has been intended for support vector machine at the application of various data mining tasks. Data Mining is a pioneering and attractive research area due to its huge application areas and task primitives. Support Vector Machine (SVM) is playing a decisive role as it provides techniques those are especially well suited to obtain results in an efficient way and with a good level of quality. In this paper, we survey the role of SVM in various data mining tasks like classification, clustering, prediction, forecasting and others applications. In broader point of view, we have reviewed the number of research publications that have been contributed in various internationally reputed journals for the data mining applications and also suggested a possible no. of issues of SVM. The main aim of this paper is to extrapolate the various areas of SVM with a basis of understanding the technique and a comprehensive survey, while offering researchers a modernized picture of the depth and breadth in both the theory and applications.

With the ever-increasing complexity and variety of database workload, database application system has been imposed on higher and higher performance requirements. Database system consists of software and hardware. And the factors that affect database performance are uncertain. In order to tackle the issue of database system for SQL server, this paper proposes hierarchical queuing network model for performance prediction, and a model is established for both software resources and hardware resources, the nested resources are linearized by hierarchical calling, thus finding out the main factors for system performance bottleneck, and system performance is adjusted and optimized accordingly. The performance tuning algorithm for SQL server database based on the hierarchical queuing network is presented in detail. And TPC-C benchmark is adopted for simulation. Experimental results show the proposed method achieves 16.8% performance increase on average, and TPS is improved by 40% compared to previous method.

In order to realize the idea of pay-as-you-go (PAYG) data management which dataspace emphasizes, the very loosely structured data model is usually used to describe the massive, heterogeneous and dynamic data in dataspace. However, the present study mainly concentrates on the applications of the data model, and the query theory research is less. The query algebra is a theoretical foundation for query and its optimization in a PAYG fashion, so that how to establish a complete algebra based on the characteristics of loosely structured data model is an important problem need to be solved. In this paper, a formal definition of very loosely structured data model is given, then the query model and query algebra based on the model are proposed, which support to not only the operations such as set operators, selection, projection and join, but also association query in dataspace.

The inherent limitations of mobile environments like disconnections, mobility, bandwidth etc affects the performance of system when several mobile clients access the shared data items at the same time. Several timeout based strategies are implemented in literature to deal with the starvation of data items during disconnections and mobility. This increases the abort and rollback rates. In the proposed heuristic approach, the time for execution of a transaction can be predicted based on the given timer value for a specific mobile client. The simulation results show the time for execution can be dynamically estimated given the timer value and vice versa for a particular request originating from specific mobile clients. This increases the throughput of the system as compared to the timer based strategies proposed in the literature.

Most of the models not aware of these dependencies on document time stamps. Not modeling time can confound co-occurrence patters and results in exchangeability of topic problem, which is important factor to deal with when finding dynamic topic discovery. This limitation has thus motivated work on developing a generalized framework for incorporating time information into topic models. Consequently, a topic model named Topics over Time (TOT) is proposed, which introduces a time node in topic model to handle the exchangeability of topics problem. However it lacks the capability to accommodate data type of side information. In this paper, we present a generative time LDA-style topic model with a variety of side information named Time Label Topic(TLT), which can find not only how the latent low-dimensional structure of document-response pairs changes over time, but also overcome the exchangeability of topics problem. Empirical results demonstrate significant improvements accuracy of time stamp and response variable prediction, and lower perplexity of our proposed model and dominance over other models.

The Entity-Relationship model (ER) is a systematic way of describing and defining many systems. The Articulated Entity Relationship (AER) is an improved model from ER-diagram by adding Functional Dependences (FDs) and normalization levels in automatically form. This paper proposed an methodology to add a new notation exactly for weak entity type and relationship between the two entity types. This notations added to the AER-diagram to facilitate the conversion from one form to another, and to make it easy to understand by the user. In addition, to it reduces the time spent by developers and designers to transform the ER-diagram to the relation (table) quickly without having to re-conversion of more than one person in the database. So, the notations define the important information in this stage. That added when convert the AAER-diagram to relation(table) in database or other form. The new notations determine the special information, such as the composite primary key for weak entity type and the primary key of the entity type that become foreign key in the other relation of other entity type according to the cardinality of the relationship. In addition, to the weak entity type and relationship between the two entity types special attributes for each one in his relation, and determined composed attribute. It be explained in detail with examples in context of this paper.

Trust Model is an efficient way of resolving the trust problems in open networks in which trust evaluation is a key issue to be addressed in trust management. According to the design rules of trust models, the problem of the lack of dynamic adapt ability in trust evaluating, the lack of effective aggregation of trust and the lack of considerations to incentive mechanisms and so on, this paper introduced the bonus-penalty factor which shows how reliable is the direct trust of the recommend entities to the subject and evaluation reliability of the recommend entities which is used to decide whether the access object would adopt the recommendation of the recommend entities during the calculation of the recommend trust. The measurement of integrated trust solves the weight problem between direct trust and recommendatory trust by introducing balanced weight factor. Finally, we present designs of bonus-penalty function and balanced weight factor and simulations by Matlab software.

The use of neural network is very wide in data mining due to some characteristic like parallel performance, Self-organizing adaptive, robustness and fault tolerance. Data mining models depend on task they accomplish: Association Rules, Clustering, Prediction, and Classification. Neural network is used to find pattern in data. The grouping of neural network model and data mining method can greatly increase the efficiency of data mining methods and it has been broadly used. Different algorithms have been discussed for optimizing the artificial neural network (ANN). ANN combines with other algorithms to find out the high accurate data as compare to traditional algorithm. The role of ANN using data mining techniques is playing an important role in forecasting or prediction about games and weather. This produces high accurate predictions than that of traditional algorithm. Data mining approaches using ANN can also work well. ANN is a highly class algorithm which can be accelerated using neuron. The result of which will produce a high speed up ANN. ANN can also be used for the purpose of extracting rules from trained neural networks.

The uncertain data as an important aspect of data mining, has received considerable attention, due to its importance in many applications, but little study has been paid to the cost-sensitive classification on uncertain data, so this paper proposes the dynamic cost-sensitive Naive Bayes classification for mining uncertain data (DCSUNB). Firstly, we apply the probability density to dispose uncertain discrete and continuous attributes, and give the cost-sensitive Naive Bayes classifier. Secondly, we propose the construction process of dynamic cost, and give the evaluation method for finding the optimal cost and the cost-sensitive classification with sequential test strategy. At last, the dynamic cost-sensitive Naive Bayes algorithm for uncertain data is structured, which searches the misclassification and test cost spaces to find the optimal cost. By comparing to the other cost-sensitive classification algorithms for uncertain data, the experiments on UCI Datasets show that DCSUNB can improve the classification performance, and reduce effectively the total cost.

In this paper we study a new method base on confidence levels for aggregating intuitionistic linguistic information. A new intuitionistic linguistic aggregation operator called the confidence intuitionistic linguistic ordered weighted averaging (CILOWA) operator is developed. Some of the CILOWA’s main properties and different families are also studied. Moreover, a practical method based on the CILOWA operator for multi-criteria decision making with intuitionistic linguistic information is presented. Finally, an illustrative example demonstrates the practicality and effectiveness of the proposed method.