Earticle

This paper is a part of ongoing research that tries to approve our hypothesis about adaptive personalization in web-based learning systems; we are debating that building online learning environments with the ability to detect learners learning style dynamically by observing their behaviour and then presenting learning material based on the detected learning style is more effective than using ILS questionnaire. In an earlier work, we provided a dynamic technique that identifies learners VAK learning styles according to their behaviour in the learning environment influenced by literature approach, in this paper we have modified our technique and re-proposed it. First, we connected behavioural patterns (time, visits and answer patterns) to the features (contents, outlines, group forums, examples, case studies, exercises and assessments), then we defined the effect each VAK learning style will have on each pattern. Next, we described three general rules and three algorithms that should detect the learning style. Now we are in the process of building two online learning environments to test and confirm the differences between dynamic learning style detection _ based on our technique _ and the traditional ILS questionnaire environments; we choose a course called ‘Computer Skills for medicine faculty students’, which is provided from the faculty of Information Technology at the University of Jordan; the results of the planed experiment will be saved and analysed using SPSS.

With cloud computing continues to mature and widely used, the availability guarantee of cloud system becomes one of the key issues. In the cloud system based on virtualization, reliability of virtual machine is an important premise to guarantee the quality of service of cloud computing. Recent research shows, applications deployed inside the virtual machine will appear performance degradation even failure after long-term operation, namely "the aging phenomenon". Prediction, diagnosis and rejuvenation of the virtual machine is a relatively complete technical framework to guarantee reliability of the cloud computing system. In this paper, we propose a software rejuvenation framework based on group migration of virtual machine to guarantee reliability of the distributed system. First construct the dependency relationship of the virtual machines; then diagnosis the software aging of virtual machine; decide the optimal virtual machine set which need to migration. Compared with the traditional single computing nodes restart/ recovery mode, downtime of group rejuvenation can be 61.53% of downtime of traditional mode. In the cloud computing environment, group rejuvenation method can ensure the availability of cloud systems more effectively than the conventional rejuvenation method.

Data mining technologies have been effectively applied to public security organizations and could help with the investigations. Among which, discovering and fighting against the gangs of criminals helps to the construction of peaceful and united environment. In this paper, we focus on the problem of discovering the gangs of criminals by integrating multiple data sources including the residential profiles from the public security bureau, the transfer data of banking accounts, the communication data from telecommunication operators, and the social interaction data from social networks. After employing a label propagation based method to discover communities in the criminal network, members within each group are ranked by their significance. Experiments exhibit the performance of proposed method.

We focus on the issue of dataset generation for fine-grained vehicle classification with CNN. Traditionally, to build a large dataset, images must be first collected manually, and then be annotated with a lot of effort. All these work are time-consuming and cost-prohibitive. In this work we propose a novel method that can generate massive images automatically, and these generated images need no annotation. An AutoCAD 3D model of a car of specified make and model is imported into our system, and then images of different views of the car are generated, these images can describe all the details of a car. By taking these images as training dataset, we use a Convolutional Neural Network to train a model for fine-grained vehicle classification. Experimental results show that these images generated virtually by 3D model indeed work as effective as real images.

Trapezoidal mechanisms and swing arm mechanisms in the steering linkages were usually designed separately, and only the steered wheel deflection errors were fixed as the optimization objective. To deal with these design faults, multi-objective optimization model of the steering linkages were developed in this paper. The model considered not only the steered wheel deflection errors but also the unevenness of the steering linkages transmitting ratios. In the model, the optimization variables were the link dimensions of the steering linkages and the optimization was processed in Adams software using SPQ algorithm. The optimization results show that the unevenness of the first axle steering mechanism transmitting ratio reduced from9.7% to 4.9%, the unevenness of the second axle steering mechanism transmitting ratio reduced from3.8% to 2.1%, the maximum deflection error of the first axle right steered wheel reduced 39.3%, the maximum deflection error of the second axle left steered wheel reduced 21.4%, the maximum deflection error of the second axle right steered wheel reduced 48.4%. Benefitting from the unevenness improvement of the steering linkages transmitting ratios, the maximum difference between the bilateral deflection errors of the first axle right steered wheel reduced from1 to 0.5 , the maximum difference between the bilateral deflection errors of the second axle left steered wheel reduced from 4 to 0.8 , the homologous maximum difference of the second axle right steered wheel reduced from 6 to 0.3 . For the handiness, the maximum difference of the bilateral steering forces reduced from1.44Nㆍm to1.1Nㆍm, the maximum difference of the bilateral number of the steering wheel turns reduced from 0.15 to 0.11. For the lemniscate simulations, the difference between the positive and negative amplitude of the yaw angular velocity reduced from1.67 / s to 0.58 / s , the difference between the positive and negative amplitude of the lateral accelerometer reduced from 2 83.3mm/ s2 to 49.6mm/ s2 . Because the unevenness of the steering linkages transmitting ratios was optimized, the steered wheel deflection errors tended to be bilateral symmetry and the vehicle handiness and control stability were both improved.

Feature reduction are common techniques that used to improve the efficiency and accuracy of the document classification systems. The problems associated with these techniques are the highly dimensionality of the feature space and The difficulty of selecting the important features for understanding the document in question. The document usually consists of several parts and the important features that more closely associated with the topic of the document are appearing in the first parts or repeated in several parts of the document. Therefore, the position of the first appearance of a word and the compactness of the word considered as factors that determine the important features using the information within a document. This study, explored the impact of combining three feature weighting methods that depend on inverse document frequency (IDF), namely, Term frequency (TFiDF), the position of the first appearance of a word (FAiDF), and the compactness of the word (CPiDF) on the classification accuracy. In addition, we have investigated different feature selection techniques, namely, Information gain (IG), Goh and Low (NGL) coefficients, Chi-square Testing (CHI), and Galavotti-Sebastiani-Simi Coefficient (GSS) in order to improve the performance for Arabic document classification system. Experimental analysis on Arabic datasets reveals that the proposed methods have a significant impact on the classification accuracy, and in most cases the FAiDF feature weighting performed better than CPiDF and TFiDF. The results also clearly showed the superiority of the GSS over the other feature selection techniques and achieved 98.39% micro-F1 value when using a combination of TFiDF, FAiDF, and CPiDF as feature weighting method.

Frequent Itemset Mining (FIM) is one of most fundamental techniques in data mining with extensive applications to a variety of data mining problems such as association rule mining, correlations, clustering and classification. Since the first proposal of frequent itemset mining, numerous serial algorithms have been proposed in order to improve mining performance, yet most of them cannot scale to massive datasets which are very common nowadays. In this paper, we propose a new parallel FIM algorithm named PFIN based on Nodeset which is a more efficient data structure for mining frequent itemsets. PFIN can intelligently decompose a large-scale FIM problem into a set of tasks, where each task can be executed in parallel without unnecessary communication overheads. Moreover, a hash-based load balancing strategy has been adopted to optimize resource use and maximize throughput. For evaluating the performance of PFIN, we have conduct extensive experiments on Spark which is an emerging distributed in-memory processing framework to compare it against PFP which is one of state-of-the-art parallel FIM algorithms on a range of real datasets. The experimental results demonstrate that our proposed PFIN are highly competitive with PFP in scalability performance, outperforming PFP in speed performance.

Event extraction is one of the most challenging tasks of information extraction from text. This paper studies one of the stages of Chinese event extraction, namely, event argument identification. A new method we call Semantic Role Labeling Based Event Argument Identification, based on the state-of-the-art methods of event extraction and event argument identification, is proposed. First, the 5W1H (who, what, whom, when, where, how) information is extracted from the text using semantic role labeling; thereafter, the 5W1H information is mapped to each argument of the event by heuristic rules. The method is used to identify the event arguments on two test sets of acquisition and transfer data, and contrasted with the methods of SRL and SRL combining heuristic rules. We find the best F1 measures for each to be 76.04% and 79.19% respectively.

Cloud computing in its various forms allows users to store their information at remote location and reduce the burden at their local systems. Even though this is an advantage for users but there are also many drawbacks because of this remote storage. The main drawback which needs to be dealt with is security. Recently, security is the major concern which most of the cloud service providers are facing. The users store their information in remote location with the hope of maintaining the privacy and integrity of data. In order, to maintain the privacy and integrity of users’ data auditing has to be done by the Cloud Service Providers (CSP). CSP uses the Third Party Auditor (TPA) for performing the auditing. The TPA performs auditing on behalf of the data owner using different auditing mechanisms. Many auditing mechanisms have been introduced in literature. Each mechanism varies from one another in one or more characteristics. In this paper we have provided a study on the different auditing mechanisms required to preserve the privacy and integrity of data in cloud. We have presented the advantages and flaws in each mechanism compared to another. Many auditing mechanisms are arising in literature with the aim to maintain the integrity of users’ data and preserve the privacy. This paper remains as the basis for different auditing mechanisms that are arising in literature. With the help of auditing mechanisms the TPA can best satisfy the needs of the users.

With the rapid development of information technology, the data scale expands unceasingly, produce a large amount of image data, a massive image data processing has also is an important technology. Traditional processing techniques, in dealing with such a large scale of image data, have been unable to meet the requirements. One of the important reason is that whether the c ++ or Java, there is memory leak problem, when the data is the relatively small size, these problems may be insignificant, but when the data is very large, will highlight these shortcomings, will lead to serious program crashes, the system outage, ultimately unable to achieve expected goal. In considering the data processing efficiency, stability, is proposed in this paper, based on multithreading, a dynamic self-healing capacity of mass image data processing technology, compared with the traditional technology can improve the efficiency and improve the stability. Finally, after the experimental results verify that the method is effective.

The most commercial search engines return the search list by matching the user query terms with the documents available in its database. The relative effectiveness of search list is highly affected by the extent to which the query keywords map to the actual need of user. User generally forms the short, ambiguous and instant queries which lead to inclusion of irrelevant documents in the search list. One well known solution to this problem is query suggestion also known as query recommendation. For query recommendations, the search systems maintain the query logs at server sites to better understand user’s information need. But till now, the current search systems have partially solved this problem as they roughly offer the similar queries to all the users regardless of their actual interests. In this paper, A novel query recommendation technique based on user browsing patterns is proposed where user interest factor in different domains are computed and used to recommend personalised queries to each individual. The experimental evaluation shows that system is able to assist user in query formation phase and efficiently reduces the search space and time required to get the desired information.

Entity recognition and entity relationship extraction are two very important tasks in information extraction. Most research work in the literature treats these two work independently when processing the text. This paper proposes a novel method for performing entity recognition and entity relationship extraction simultaneously from unstructured text based on Conditional Random Fields (CRFs). This method makes use of entity features, entity relationship features and features of the triples which is composed of entities and their relationship to conduct the model training. Experiment results show that this method can recognize entity and extract entity relationship effectively.

Context: Global Software Development (GSD) is the software development across the globe in which stakeholders are related to different countries and cultures, and they communicate each other by emails, fax, mobile, videoconferencing or any other communicating media. There are a lot of problems in implementing requirement engineering process for global software development. There is a need of requirement implementation model which guides us how to implement successfully requirement engineering in the context of GSD. Objective: To Find Critical Success factors (CSF) and challenges in requirement implementation in the context of GSD through Systematic literature Review (SLR), and to find the practices for the proper implementation CSF and challenges as proposed by Requirement Implementation Model (RIM). Method: - SLR is the methodology used to fulfill the objectives of this research. Expected Outcomes: - SLR protocol is developed for RIM. Expected output of this study is to list out all the factors and challenge which the stakeholders are facing in implementing requirement in the context of GSD through SLR.

As the network promoting in people's lives, more and more people use the social networks and other network platform, the concepts of big data have been gradually reference to the network data analysis, researching big data analysis framework and technology can improve enterprise management level. This paper analyzes the framework and its key technology of big data analysis in the context of social network, in order to be able to provide the corresponding theoretical support and reference for analysis of social network data.

With the development of society, number of art image becomes bigger and bigger. In order to classify the art images, it is necessary to retrieve the images with common features. Different from the commonly used image query method, art images are always seen by specific researchers and to study and classification, and it is much more important for the query method with high retrieval precision. So it is important to develop a better query method specifically used for the art image query. The main work of this paper is to establish a combined method to improve the art image query precision without decreasing the query time seriously. The combined query method includes several steps: (1) Initial query. The tag query method and the semantic query method will be used. The initial query results includes the results searched by both the methods. (2) Reorganization of the initial query results. The repeated images will be cleared and only one image will be left. (3) Distance based results rechecked. The most relative image will be ranked in front of the list and the less correlation image will be cleared. (4) Reordering the images. All the images will be reordered by the distance which reflects the correlation of the query image. According to the experimental verification, the combined method can improve the query precision.

Machine learning utilizes algorithms to run predictive models that learn from a large dataset in an iterative manner. Predictive models are used in many business applications to gain competitive advantages and understand customers better. This paper concentrates on analyzing New York taxi trips and fares and presenting the methodology we used to address the problem and results reached by building through Azure Machine learning studio. Our practical approach starts with an exploratory analysis of NYC taxi data via Microsoft Power BI. Then more extensive analysis was conducted through Apache Hive data warehouse. Hive was built on top of Hadoop enabling data synopsis, query, and analysis. We implemented Hive queries to create tables in Microsoft Azure blob storage and store the data in external tables. Finally, we conducted our experiment by creating, training and testing the module. The finding and insights pertain to the main variables of our experiment: pick up time, drop off time and tip amount that could be integrated into an application and enhance business by picking the location with the highest tip for example.

Art image databases play an important role in the image research history. Art image can be classified into various categories, and each category has its own characteristics. Art image retrieval with high precision and speed can help the researchers understand an art image much more easily. In addition, the classification for the art images can help the researchers improve the working efficiency. In the art image retrieval, the query precision is much more important than the query speed. Improving the query precision at the cost of not seriously decreasing the query speed can be accepted. In this paper, a new method has been proposed to improve the query precision. The new method mainly includes initial query, reorganization, results recheck and images reordering. At the beginning, the tag query method and the semantic query method will be used to search the initial image results; then, the results will be reorganization according to the semantic method; finally, the images in the results will be filtered by the incomplete filtration method. According to the experimental results, the new method is proved that it can improve the query precision. The new method can be used in the art image retrieval process.

Multi Query Optimization is one of the most important tasks in Relational Database Management System (RBMS) and it becomes common due to high usage of online decision support management systems in every industry nowadays. In multi query optimization, queries are optimized and executed in batches. However, there are many algorithms use to detect and unified common sub-expressions among multiple queries and unified them so that the more encompassing sub- expression is executed and the other sub-expressions are derived from. In this work, multi-query optimization algorithm using heuristics and semantic approaches was proposed and encoded on SQL Server version 10.0.1600 and three queries were used for the experiment between the proposed algorithm and most recent basic Multi Query Optimization Algorithm (Volcano RU). The result of experiment showed that, Proposed Algorithm gave the best plans compared Volcano RU Algorithm, across all three queries and was best for all queries in terms of execution time and CPU time.

Traditional affinity propagation algorithm has inefficient results when conducting clustering analysis of high dimensional data because "dimension effect" lead to difficult find the proper class structure .In view of this, the author proposes an improved algorithm on the basis of Entropy Weight Method and Principal Component Analysis (EWPCA-AP). EWPCA-AP algorithm empowers the sample data by Entropy Weight Method, eliminate data irrelevant attributes by Principal Component Analysis, and travel with neighbor clustering algorithm, realization of high-dimensional data clustering in low dimension space. The numerical result of simulation experiment shows that the new EWPCA-AP algorithm can effectively eliminate the redundancy and irrelevant attributes of data and improve the performance of clustering. In addition, the proposed algorithm is applied in the area of the economy in our country and the clustering result is consistent with the real one. This algorithm provides a new intelligent evaluation method for Chinese economy.

Given the performance of original affinity propagation algorithm is greatly affected by preference (P), stability threshold-based affinity propagation clustering algorithm (STAP) is proposed in this paper, including stability threshold to obtain the state of convergence when getting real class number and capture the corresponding P, and it take S-type function as damping factor to accelerate the convergence speed of STAP clustering algorithm. Besides it is successfully applied in the financial evaluation of public companies. The simulation experimental results show that, comparing the traditional affinity propagation clustering algorithm, STAP clustering algorithm can obtain high precision and fast convergence rate to improve clustering performance.

World Wide Web contains huge amount of data available in different languages across the world. Web browsers are tools used to display the data in graphical forms. With the evolution of Web 3.0, data has become an important part of human daily tasks, where it is used to process information, and formulate important decision rules for many organizations. Current tools used to conceptualize data are catered for some of the world well known languages such as English. However, these tools may not be able to support other languages as there are a wide range of languages with different syntax and representation. In this paper, we present a novel lexical semantic based database tool called MalayWordNet, specifically written for Malay language. Our tool is helpful for high end semantic based applications which use Malay language as part of their data presentation.

Object: With the continuous advancement of information technology, more and more systems use database to store basic data, the security of data is an important part of the design of the business system. Method: In order to effectively improve the viability of data, this paper proposes a kind of multi point and multi hop database remote disaster recovery and backup technology. Process: Based on the in-depth analysis of the functions and demand of database, this paper introduces the working principles and key technologies of disaster recovery technology, describes the principles and realization process of multi point and multi hop database remote disaster recovery and backup technology, and carries on the experimental analysis. Conclusion: Theoretical analysis and experimental results show that this method is an effective new way of database remote disaster recovery and backup. So this technology has many advantages, such as multi point and multi hop backup, good real-time performance, fine backup granularity and so on.

Under the highly development of the information society, only by carrying out education reform can be cultivate more innovative talents. Because the traditional education idea has already taken root in the hearts of people, education informatization is a necessary way to change this kind of thought. The generation of cloud computing technology leads to the revolution of data processing technology. It can make use of a small amount of resources to effectively deal with the big data in the information system of educational institutions. Neural network is one of the important technologies of educational data mining in cloud computing environment. BP neural network is a typical multi-layer forward network, which is composed of input layer, hidden layer and output layer. It can be used to predict the data through the training model. In this paper, based on the characteristics of the distribution of education resources, we put forward the method to analyze big data of education by using Hadoop technology. This method uses the MapReduce programming model to manage the data, so as to improve the speed and efficiency of data analysis. Secondly, in Hadoop platform, this paper puts forward the method of parallel BP neural network in education data processing. The method consists of the following main steps: firstly, input data and set up a three layer parallel neural network. Secondly, according to the location of each node to block the data, and transfer M separate blocks to the Map function for processing. Thirdly, through the gradient descent method, the Map function finds the weight distribution of each block by iterative algorithm. Fourthly, we transfer the key-vlaue to the Reduce function, and update the statistics. Finally, repeat the update the calculation process of weight. After several iterations, the optimal solution of the objective function is found, and the weight distribution of the network is obtained. Finally, we simulate the parallel BP neural network algorithm based on education cloud platform, in order to prove that it is suitable for the prediction of learning achievement of the network teaching system.

Due to its importance in many applications, the incomplete data mining has received increasing attention in recent years, but there has been little study of the cost-sensitive classification on incomplete data. Therefore this paper proposes the dynamic cost-sensitive extreme learning machine for classification of incomplete data based on the deep imputation network (DCELMIDC). Firstly, we propose an approach for incomplete data imputation based on the deep imputation network model, and offer the cost-sensitive extreme learning machine. Secondly, this paper introduces dynamic misclassification and test cost, and gives the chromosome coding and an evaluation method of the optimal cost. At last, on the basis of the genetic algorithm, the dynamic cost-sensitive extreme learning machine classification algorithm for mining incomplete data is given, which can search the optimal misclassification and test cost in cost spaces. The experiment results show that DCELMIDC is effective and feasible for classification of incomplete data, and can reduce the total cost.