Traditional association-rule mining (ARM) considers only the frequency of items in a binary database, which provides insufficient knowledge for making efficient decisions and strategies. The mining of useful information from quantitative databases is not a trivial task compared to conventional algorithms in ARM. Fuzzy-set theory was invented to represent a more valuable form of knowledge for human reasoning, which can also be applied and utilized for quantitative databases. Many approaches have adopted fuzzy-set theory to transform the quantitative value into linguistic terms with its corresponding degree based on defined membership functions for the discovery of FFIs, also known as fuzzy frequent itemsets. Only linguistic terms with maximal scalar cardinality are considered in traditional fuzzy frequent itemset mining, but the uncertainty factor is not involved in past approaches. In this paper, an efficient fuzzy mining (EFM) algorithm is presented to quickly discover multiple FFIs from quantitative databases under type-2 fuzzy-set theory. A compressed fuzzy-list (CFL)-structure is developed to maintain complete information for rule generation. Two pruning techniques are developed for reducing the search space and speeding up the mining process. Several experiments are carried out to verify the efficiency and effectiveness of the designed approach in terms of runtime, the number of examined nodes, memory usage, and scalability under different minimum support thresholds and different linguistic terms used in the membership functions.
In this paper we deal with misbehaving nodes in mobile ad hoc networks (MANETs) that drop packets supposed to be relayed, whose purpose may be either saving their resources or launching a DoS attack. We propose a new solution to monitor, detect, and safely isolate such misbehaving nodes, structured around five modules: (i) The monitor, responsible for controlling the forwarding of packets, (ii) the detector, which is in charge of detecting the misbehaving of monitored nodes, (iii) the isolator, basically responsible for isolating misbehaving nodes detected by the detector, (iv) the investigator, which investigates accusations before testifying when the node has not enough experience with the accused, and (v) finally the witness module that responds to witness requests of the isolator. These modules are based on new approaches, aiming at improving the efficiency in detecting and isolating misbehaving nodes with a minimum overhead. We describe these modules in details, and their interactions as well. We also mathematically analyze our solution and assess its performance by simulation, and compare it with the watchdog, which is a monitoring technique employed by almost all the current solutions. 相似文献
When focusing on the general area of data mining, high-utility itemset mining (HUIM) can be defined as an offset of frequent itemset mining (FIM). It is known to emphasize more factors critically, which gives HUIM its intrinsic edge. Due to the flourishing development of the IoT technique, the uncertainty patterns mining is also attractive. Potential high-utility itemset mining (PHUIM) is introduced to reveal valuable patterns in an uncertainty database. Unfortunately, even though the previous methods are all very effective and powerful to mine, the potential high-utility itemsets quickly. These algorithms are not specifically designed for a database with an enormous number of records. In the previous methods, uncertainty transaction datasets would be load in the memory ultimately. Usually, several pre-defined operators would be applied to modify the original dataset to reduce the seeking time for scanning the data. However, it is impracticable to apply the same way in a big-data dataset. In this work, a dataset is assumed to be too big to be loaded directly into memory and be duplicated or modified; then, a MapReduce framework is proposed that can be used to handle these types of situations. One of our main objectives is to attempt to reduce the frequency of dataset scans while still maximizing the parallelization of all processes. Through in-depth experimental results, the proposed Hadoop algorithm is shown to perform strongly to mine all of the potential high-utility itemsets in a big-data dataset and shows excellent performance in a Hadoop computing cluster.
This paper presents an asynchronous cascading wake-up MAC protocol for heterogeneous traffic gathering in low-power wireless sensor networks. It jointly considers energy/delay optimization and switches between two modes, according to the traffic type and delay requirements. The first mode is high duty cycle, where energy is traded-off for a reduced latency in presence of realtime traffic (RT). The second mode is low duty cycle, which is used for non-realtime traffic and gives more priority to energy saving. The proposed protocol, DuoMAC, has many features. First, it quietly adjusts the wake-up of a node according to (1) its parent’s wake-up time and, (2) its estimated load. Second, it incorporates a service differentiation through an improved contention window adaptation to meet delay requirements. A comprehensive analysis is provided in the paper to investigate the effectiveness of the proposed protocol in comparison with some state-of-the-art energy-delay efficient duty-cycled MAC protocols, namely DMAC, LL-MAC, and Diff-MAC. The network lifetime and the maximum end-to-end packet latency are adequately modeled, and numerically analyzed. The results show that LL-MAC has the best performance in terms of energy saving, while DuoMAC outperforms all the protocols in terms of delay reduction. To balance the delay/energy objectives, a runtime parameter adaptation mechanism has been integrated to DuoMAC. The mechanism relies on a constrained optimization problem with energy minimization in the objective function, constrained by the delay required for RT. The proposed protocol has been implemented on real motes using MicaZ and TinyOS. Experimental results show that the protocol clearly outperforms LL-MAC in terms of latency reduction, and more importantly, that the runtime parameter adaptation provides additional reduction of the latency while further decreasing the energy cost. 相似文献
The challenging problem of time synchronization in wireless sensor networks is considered in this paper, where a new distributed protocol is proposed for both local and multi-hop synchronization. The receiver-to-receiver paradigm is used, which has the advantage of reducing the time-critical-path and thus improving the accuracy compared to common sender-to-receiver protocols. The protocol is fully distributed and does not rely on any fixed reference. The role of the reference is divided amongst all nodes, while timestamp exchange is integrated with synchronization signals (beacons). This enables fast acquisition of timestamps that are used as samples to estimate relative synchronization parameters. An appropriate model is used to derive maximum likelihood estimators (MLE) and the Cramer-Rao lower bounds (CRLB) for both the offset-only, and the joint offset/skew estimation. The model permits to directly estimating relative parameters without using or referring to a reference’ clock. The proposed protocol is extended to multi-hop environment, where local synchronization is performed proactively and the resulted estimates are transferred to the intermediate/end-point nodes on-demand, i.e. as soon as a multi-hop communication that needs synchronization is initiated. On-demand synchronization is targeted for multi-hop synchronization instead of the always-on global synchronization model, which avoids periodic and continuous propagation of synchronization signals beyond a single-hop. Extension of local MLE estimators is proposed to derive relative multi-hop estimators. The protocol is compared by simulation to some state-of-the-art protocols, and results show much faster convergence of the proposed protocol. The difference has been on the order of more than twice compared to CS-MNS, more than ten times compared to RBS, and more than twenty times compared to TPSN. Results also show scalability of the proposed protocol concerning the multi-hop synchronization. The error does not exceed few microseconds for as much as 10 hops in R4Syn, while in CS-MNS, and TPSN, it reaches few tens of microseconds. Implementation and tests of the protocol on real sensor motes confirm microsecond level precision even in multi-hop scenarios, and high stability (long lifetime) of the skew/offset model. 相似文献
This paper proposes a novel framework for metaheuristic-based Frequent Itemset Mining (FIM), which considers intrinsic features of the FIM problem. The framework, called META-GD, can be used to steer any metaheuristics-based FIM approach. Without loss of generality, three metaheuristics are considered in this paper, namely the genetic algorithm (GA), particle swarm optimization (PSO), and bee swarm optimization (BSO). This allows to derive three approaches, named GA-GD, PSO-GD, and BSO-GD, respectively. An extensive experimental evaluation on medium and large database instances reveal that PSO-GD outperforms state-of-the-art metaheuristic-based approaches in terms of runtime and solution quality. 相似文献