Huafeng Wu

Add Semantic Role to Dependency Structure Language Model for Topic Detection and Tracking

In this paper, an idea of adding semantic role to the dependency structure language model is proposed. Firstly, the dependency structure language model for topic detection and tracking is presented. Then we introduce the method to determine the semantic role for the constituents of a sentence. Finally, we add the semantic role to the dependency structure language model Compare the verbs of the sentences in the stories with a list of verbs related with the verb of the topic. Then, annotate the verbs with semantic roles. This can enable us establish a relation between topics and semantic roles. So, only stories whose sentences containing the right semantic roles are selected. We propose using this semantic information as an extension of the dependency structure language model in order to reduce the number of stories retrieved by the system, and get a high precision in topic detection and tracking.

Reputation-based Trust in Wireless Sensor Networks

Wireless sensor networks (WSNs) are becoming more and more common. WSNs have many existing and envisioned applications due to their ease of deployment, particularly in remote areas. But the security of WSNs is still an issue. Some existing approaches mainly rely on cryptography to ensure data authentication and integrity. These approaches only address part of the problem of security in WSNs. However, these approaches are not sufficient for the unique characteristics and novel misbehaviors encountered in WSNs. Recently, the use of reputation systems has become an important mechanism in WSNs. In this paper we propose a reputation-based trust which borrows tools from probability, statistics and mathematics analysis. We have suggested a new term certainty used in trust system and we argued that the positive or negative outcomes for a certain event is not enough information to make a decision in WSNs. We build up a reputation space and trust space in WSNs, and define a transformation from reputation space to trust space. Finally, we discuss some important properties of them and point out some open problems in reputation system in WSNs.

Agent-Based Trust Management Model for Wireless Sensor Networks

In wireless sensor networks (WSNs), these sensor nodes cooperate with each other to form a network without using any infrastructure. WSNs has a wide application. But the security of WSNs is still an issue. Traditionally cryptographic mechanisms such as authentication and encryption only address part of the problem of security in WSNs. Recently, the use of reputation and trust systems has become an important mechanism in WSNs. In this paper, we develop a distributed agent-based trust management scheme. The agent node use watchdog mechanism to observe the behavior of the sensor nodes and computes the trust rating for them. We consider packet-dropping and hello flood attack in our simulation, the simulation results and analysis show that our scheme can fast detect the malicious nodes and scale well for wireless sensor networks.

Event-based Trust Framework Model in Wireless Sensor Networks

The security of wireless sensor networks is ever more important nowadays. Most of the proposed security protocols in wireless sensor networks are based on authentication and encryption. But all of them only address part of the problem of security in wireless sensor networks. Recently, the use of reputation and trust systems has become an important secure mechanism in wireless sensor networks. In this paper we propose a new protocol ETSN to construct a trust framework model in wireless sensor networks. The ETSN will be more suitable than ATSN and RFSN in wireless sensor networks. The simulation results and analysis show that our scheme not only can fast detect the malicious nodes and scale well for wireless sensor networks, but also can distinguish trust rating with different event.

Trust Propagation and Aggregation in Wireless Sensor Networks

In a multi-hop wireless sensor networks (WSNs), these sensor nodes cooperate each other to form a network without using any infrastructure, which has a wide application due to their easy of deployment. But the security of WSNs is still an issue. Traditionally cryptographic mechanisms such as authentication and encryption only address part of the problem of security in WSNs. Recently, the use of reputation systems has become an important mechanism in WSNs. The trust management is still a very important problem for reputation systems in WSNs. In this paper we develop a formal treatment of trust management for WSNs. The main contribution of our paper is that we define two kinds of operators, propagation and aggregation. It can combine and diffuse trust rating of nodes in WSNs. Simulation results and analysis show that our proposed operators can detect the misbehavior sensor fast and limited the effect of malicious report the trust rating.

Throughput Performance of Marine STDMA Ad-hoc Network

Due to the convenience of Internet, people can search for whatever information they need and buy whatever they want on the web. In the age of E-Commerce, it is difficult to provide support for customers to find the most valuable products that match their heterogeneous needs. Traditional approaches to this so-called personalization problem adopt predefined formats to describe the customer requirements. This always leads to distortion in eliciting requirement information and thus inaccurate recommendations. In this paper, we propose a personalized recommendation system using association rule mining and classification in e-commerce. Customer requirements are extracted from text documents and transformed into a set of significant phrases. Allowing the transformed transaction records, a set of association rule are mined from database using Apriori algorithm. CBA-CB algorithm is applied to produce the best rules out of the whole set of rules. The best classifiers are then generated after the test and validation of those rules, aimed to predict the item labels for new customer requirements and thus assigns the corresponding class labels to the customer. The system analysis and design of the proposed recommendation system as well as the implementation of prototype are also presented.The spatial reuse time division multiple access (STDMA) ad hoc network got applied in the marine navigation realm. The resulted system, automatic identification system (AIS), has been extensively used in marine industry. It well serves the surveillance and communication purposes for marine navigation, and makes a great improvement on marine technology. Throughput is an important measure to such a surveillance and communication system, especially with the rapid increase of the amount of ships and their speed. In this paper, we firstly find out the influence factors of throughput through analysis of the systems basic principle. Through simulation, we simulate and analyze how these factors impact the throughput performance. Finally, we simply propose some suggestions for improving it within different environment.

A rank-based application-driven resilient reputation framework model for wireless sensor networks

The security of wireless sensor networks is an important field of research about WSNs, and has attracted a lot of attentions nowadays. Reputation based trust systems represent soft security mechanisms that complement traditional information security. Nowadays, the research of trust security mechanism based on reputation and has become a hot research issue in wireless networks. In this paper, we propose a new security protocol RARRM to construct trustworthy sensor networks. This protocol not only can distinguish different applications, but also can rank trust values. The objective of this protocol is not only to encourage nodes to provide good services consecutively to establish a security environment of network, but also to punish strictly those malicious nodes that are trying to take advantage of reputation system to launch malicious attacks. Through intensive simulation, we will verify the correctness and efficiency of the proposed defense techniques.

Key Feature and Rule-based Intrusion Detection for Wireless Sensor Networks

Wireless sensor networks (WSNs) have a wide application. However, they are of interest to adversaries and are susceptible to some types of attacks for they are deployed in unprotected environments. In this paper we use key feature and rule-based intrusions detection for WSNs. In our scheme, we investigate different defense rules and key features in WSNs to build efficient, accurate and effective intrusion detection systems (IDSs). We also propose a moving window function method to gather the current activity data. The scheme fits the demands and restrictions of WSNs. The scheme does not need any cooperation among other monitor nodes. Simulation results show that we proposed IDSs is efficient and accurate.