Chanle Wu

Solving the Sparsity Problem in Recommender Systems Using Association Retrieval

Recommender systems are being widely applied in many fields, such as e-commerce etc, to provide products, services and information to potential customers. Collaborative filtering as the most successful approach, which recommends contents to the current customers mainly is based on the past transactions and feedback of the similar customer. However, it is difficult to distinguish the similar interests between customers because the sparsity problem is caused by the insufficient number of the transactions and feedback data, which confined the usability of the collaborative filtering. This paper proposed the direct similarity and the indirect similarity between users, and computed the similarity matrix through the relative distance between the user’s rating; using association retrieval technology to explore the transitive associations based on the user’s feedback data, realized a new collaborative filtering approach to alleviate the sparsity problem and improved the quality of the recommendation. In the end, we implemented experiment based on Movielens data set, the experiment results indicated that the proposed approach can effectively alleviate the sparsity problem, have good coverage rate and recommendation quality.

Flexible Advance Reservation for Grid Computing

Requests of advance reservation with fixed parameters, i.e. start time, end time and resource capability, may be rejected due to instantaneous peaks of resource utilization. Gaps between these peaks are too narrow for additional requests to fit in. As a result, the call acceptance rate of reservation would decrease dramatically, and the performance of resource may be reduced. In fact, many resource reservations for grid applications don’t need fixed parameters. In this paper, a flexible advance reservation is introduced. Its parameters can be modified according to resource status in order to fill the gaps of resource. Particular admission control algorithm for this new type of reservation is provided too. Simulation shows that it can improve performance of resource reservation in terms of both call acceptance rate and resource utilization.

Packet Scheduling Using Bidirectional Concurrent Transmission in WiMAX Mesh Networks

Performance of wireless mesh networks can be greatly improved by employing spatial reuse with concurrent transmissions. However, most current research works on centralized scheduling for WiMAX mesh networks are based on a unidirectional concurrent scheme, i.e., transmissions of uplinks and downlinks are considered separately. In this paper, we analyze the interference model of IEEE 802.16 TDMA mesh networks and propose a bidirectional concurrent transmission model. We propose a packet scheduling problem to formulate centralized scheduling in WiMAX mesh networks with bidirectional concurrent transmission, and make simulations to verify the effect of it with various scheduling algorithms. Experiment results reveal that comparing to unidirectional transmission, bidirectional transmission can effectively minimize the number of timeslots required to transmission packets.

Research on an Effective Mechanism of Task-scheduling in Grid Environment

This paper gives a definition of the communication overhead among nodes in task scheduling and resource management in grid environment, presents rules of building QoS routing tree based on the QoS routing algorithm, proposes grid heap-sort algorithm and QoS routing algorithm, and designs a scheme of grid task scheduling with the help of them. The results of its correlative experiment indicate that this design can guarantee quality of service in grid better

Balanced and resilient application layer multicast

In the context of Application Layer Multicast(ALM), quality of high-bandwidth real-time communication over best effort networks is mainly constrained by node capacity and link loss characteristics observed along multicast delivery path. Our ALM solution is based upon Cluster-Oriented, Regular and Multi-trees(CORM) structure and can overcome these constraints by adequately supporting loss recovery, load-balance and moreover, loss-balance (the loss caused by nodes failures is roughly balanced and within a small range). Clusters in a hierarchy uniformly distribute forwarding load among all participating peers. Regular (constant node degree) constraint makes cluster maintenance and optimization as a whole. Furthermore, for improving robustness, we design and implement linear network coding scheme, proactive parent recovery mechanism and cluster splicing algorithm in terms of CORM. Experimental results obtained via network simulation show that CORM has lower average link stress (by about 22%), and more even node stress than NICE protocol. More importantly, failure recovery properties such as recovery time and loss-balance, are much better by using orders of magnitude lower control traffic. Moreover, we analyze performance effect caused by combining various regular degree and cluster size.

Throughput Enhancement with Bidirectional Concurrent Transmission in IEEE 802.16 Mesh Networks

Performance of wireless mesh networks can be greatly improved by spatial reuse with concurrent transmissions. However, most current research works on WiMAX mesh networks are based on a unidirectional concurrent transmission scheme to perform link and packet scheduling. In this paper, we analyze the interference model of IEEE 802.16 TDMA mesh networks and propose a bidirectional concurrent transmission model. We formulate the model into an ILP problem and provide the solution. Simulation results show that our bidirectional transmission mechanism can effectively improve the network performance. Further study reveals that the throughput enhancement varies with different network topologies and distinct location of BS.

Semantic Learning Service Personalized

Abstract To provide users with more suitable and personalized service, personalization is widely used in various fields. Current e-Learning systems search for learning resources using information search technology, based on the keywords that selected or inputted by the user. Due to lack of semantic analysis for keywords and exploring the user contexts, the system cannot provide a good learning experiment. In this paper, we defined the concept and characteristic of the personalized learning service, and proposed a semantic learning service personalized framework. Moreover, we made full use of semantic technology, using ontologies to represent the learning contents and user profile, mining and utilizing the friendship and membership of the social relationship to construct the user social relationship profile, and improved the collaboration filtering algorithm to recommend personalized learning resources for users. The results of the empirical evaluation show that the approach is effectiveness in augmenting ...

Ant colony optimization for PON network design

Optical access technology has quickly got a worldwide development. Currently, the deployment of equipment and fiber lines in the optical access network require complex manual design. However, because of non-standard artificial design, it is difficult to make the network reasonable, and increasing post-engineering and maintenance difficulties. Based on the partition algorithm in the previous article, this paper presents a new partitioning algorithm for the PON network design, which is a combination of the Ant Colony Optimization algorithm and partitioning algorithm. Compared with other algorithms, the ant colony optimization is more efficient and more accurate. It is very suitable for solving calculating the capacity of the large and complex network. Therefore, we use ant colony optimization for PON network design. By entering the number of nodes, users, service traffic, the unit cost parameters, it can automatically calculate the number and configuration of OLT, which can help to optimize the network layout, and save the network construction costs.

A Connectivity-Prior Model for Generating Connected Power Law Random Graphs with Prescribed Degree Sequence

Generating precise network topologies is an important issue for the purpose of simulating and evaluating networking applications. Recent research results reveal that the topology of Internet is neither a purely random network nor a hierarchical structure, but similar to complex networks obeying power law distributions. Under this condition, a practical degree-driven method is widely used for generating network topologies with prescribed degree sequence. To import random features, additional random transformations are required to perform upon the generated graph. In this paper, we propose a connectivity-prior algorithm to create a connected graph and develop a simple but efficient method to perform randomization operations to transform the generated graph. During the creating and transforming process, the graph is kept connected. We made experiments with the latest degree sequence data of the actually Internet topologies. The results show that our method works more efficiently.

Personalized learning service framework based on Semantic Web Technologies

Personalization is widely used in various fields to provide users with more suitable and personalized service. In this paper, a personalized learning service framework based on Semantic Web Technologies is proposed. The framework made full use of ontology to construct the semantic metadata model and user profile, mining and utilizing the friendship and membership to provide personalized learning services that include learning service and learning support service by a new ontology-based collaborative recommendation algorithm. Experiments in this paper showed that the proposed architecture is effective.