Abstract Efficient and accurate sales forecasting is a vital part of creating an efficient supply chain in enterprises. Times series methods are a popular choice for forecasting demand sales. A major challenge is to develop a relatively inexpensive and automated forecasting engine that guarantees a desired forecasting accuracy. Times series decomposition and Forecast combination have been two classes of methods that... Efficient and accurate sales forecasting is a vital part of creating an efficient supply chain in enterprises. Times series methods are a popular choice for forecasting demand sales. A major challenge is to develop a relatively inexpensive and automated forecasting engine that guarantees a desired forecasting accuracy. Times series decomposition and Forecast combination have been two classes of methods that have attracted the interest of recent researchers. One solution has been to use decomposition followed by recombining to form a very large number of forecasting models Further many recent papers present Data Mining based methods to intelligently discover a subset of methods from a large list that can be used in combination for sales demand forecasting. These methods are computationally expensive and prohibitive if they are applied to each individual time series in a retail organization. In this paper we present a novel technique to identify similar sales series and efficiently use the best combination of methods learnt for one series to forecast for the entire set of similar series. Read More Read Less

Abstract In order to improve the accuracy of price forecasting for dynamic trend analysis on imperfect data by web extracting, a novel repair data algorithm based on dichotomy backfilling is proposed in this paper. The price forecasting algorithm based on the Sliding Window (SW) is utilized to verify the validity of Repair Dichotomy Backfilling (RDB) algorithm. Experiments demonstrated that the mean... In order to improve the accuracy of price forecasting for dynamic trend analysis on imperfect data by web extracting, a novel repair data algorithm based on dichotomy backfilling is proposed in this paper. The price forecasting algorithm based on the Sliding Window (SW) is utilized to verify the validity of Repair Dichotomy Backfilling (RDB) algorithm. Experiments demonstrated that the mean absolute errors can be achieved to 5.31 percent average. Furthermore, the price forecasting algorithm based on original data and that based on RDB data are explored to compare their accuracy. Experiments demonstrated the reduction of mean absolute errors in the price dynamic trend analysis verification model. The average MAE on original data is 2.38 percent when the average MAE on RDB data is 2.31 percent. Experiment results proved that the proposed RDB-SW approach is meaningful and useful to analyze and to research the price market on imperfect data by Web extracting. Read More Read Less

Abstract Live P2P streaming has been widely used in recent years because of its scalabil-ity and low costs. At the same time, it meets some new challenges that centralized systems never faced, for example, churn and content pollution. It is the first paper which uses stochastic activity networks (SANs) to model and analyze pushbased live P2P streaming systems from the users... Live P2P streaming has been widely used in recent years because of its scalabil-ity and low costs. At the same time, it meets some new challenges that centralized systems never faced, for example, churn and content pollution. It is the first paper which uses stochastic activity networks (SANs) to model and analyze pushbased live P2P streaming systems from the users point of view. The model provides a new line of thinking to model push-based live P2P streaming, which enables the effect of churn, pollution and buffer size to be studied simultaneously. It explains why there is a gap between theoretical results in previous studies of content pollution in P2P streaming and the observed effect. It also points out the threshold of buffer size for non-significant QoS improvement. Read More Read Less

Abstract The success of Linked Data project has played a vital role in the realization of the Semantic Web on a global stage. It has motivated people to publish datasets which are important for information linking, resource discovery and can further make contributions in shaping the Web as a single connected data space. This effort has successfully amassed a variety of... The success of Linked Data project has played a vital role in the realization of the Semantic Web on a global stage. It has motivated people to publish datasets which are important for information linking, resource discovery and can further make contributions in shaping the Web as a single connected data space. This effort has successfully amassed a variety of Linked Data and has introduced many novel ways for the publishing of data. As a result, putting Linked Data online has become rather easy, but actually linking the data with already existing data in the cloud and further on presentation of post-interlinking data are still researchable challenges. The search and identification of relevant datasets as well as devising a strategy for linking to these datasets with post-organization of interlinked data for human perceivable presentation is still a difficult task. In this paper, a novel approach is presented which 1) highlights and implements the steps involved in the interlinking process and 2) proposed a proof of concept application for automatic generation of an organized profile for post interlinked data presentation. These approaches are applied and presented as a case study focusing on interlinking scholarly communication datasets and highlighting the potential benefits of Linked Data. This study has interlinked two semantified scholarly datasets and was successfully able to discover new resources which were further organized into profile for showcasing potentials of Linked Data. Read More Read Less

Abstract With the advance of cloud computing, the trend of developing and deploying applications on cloud environment also has appeared recently. There are economic as well as technological reasons why an application should be developed and deployed on the cloud. Reducing costs and business agility are typical business factors for cloud-based applications. Cloud computing can provide significant cost savings because of... With the advance of cloud computing, the trend of developing and deploying applications on cloud environment also has appeared recently. There are economic as well as technological reasons why an application should be developed and deployed on the cloud. Reducing costs and business agility are typical business factors for cloud-based applications. Cloud computing can provide significant cost savings because of the increased utilization resulting from the pool-ing of resources. In addition, this technology enables rapid delivery of IT services, which also improves business efficiency. On the operational side, manageability, performance, and scalability are the typical reasons why application developers consider cloud computing. However, cloud computing is still considerable complex to use. Not only the migration between clouds is difficult - but also the development and deployment of a new, complex service from the beginning could be quite a challenge using today's cloud computing. In this paper, novel high-level abstraction layers for cloud computing is presented. The abstraction layers will allow users to manage cloud resources form various clouds and simplify the process of developing and deploying services into those clouds. This approach will also solve absolutely the interoperability issue, thus improving the flexibility of cloud computing. Read More Read Less

Abstract Trust evaluation is an essential service to ensure reliability on security relationships on networks, including Mobile Ad Hoc Networks (MANETs). However, the implementation of an effective trust evaluation service is very difficult in these networks, due to their dynamic characteristics. Though many trust evaluation services for MANETs can be found in the literature, they were not designed considering malicious attacks... Trust evaluation is an essential service to ensure reliability on security relationships on networks, including Mobile Ad Hoc Networks (MANETs). However, the implementation of an effective trust evaluation service is very difficult in these networks, due to their dynamic characteristics. Though many trust evaluation services for MANETs can be found in the literature, they were not designed considering malicious attacks or were not evaluated under these attacks. This work presents TRUE, a distributed trust evaluation service for MANETs which creates a self-organized context-based trust network. To estimate the trustworthiness of other nodes, nodes form trust chains based on behavior evidences maintained within the context-based trust network. Periodically, nodes exchange trust information with the neighbors, providing an efficient method to disseminate these information through the network. Simulation results show that TRUE is very efficient on gathering evidences to build the trust networks, maintaining a small communication and memory overhead. Further, it is the first trust evaluation scheme evaluated under bad mouthing and newcomers attacks and it maintains its effectiveness in such scenarios. Read More Read Less