Intrusion detection systems help network administrators prepare for and deal with network security attacks. These systems collect information from a variety of systems and network sources, and analyze them for signs of intrusion and misuse. A variety of techniques have been employed for analysis ranging from traditional statistical methods to new data mining approaches. In this study the performance of three data mining methods in detecting network intrusion is examined. An experimental design (3times2x2) is created to evaluate the impact of three data mining methods, two data representation formats, and two data proportion schemes on the classification accuracy of intrusion detection systems. The results indicate that data mining methods and data proportion have a significant impact on classification accuracy. Within data mining methods, rough sets provide better accuracy, followed by neural networks and inductive learning. Balanced data proportion performs better than unbalanced data proportion. There are no major differences in performance between binary and integer data representation. 相似文献
Journal of Combinatorial Optimization - This paper proposes an online leasing problem considering both price fluctuations and the second-hand transaction. In the studied problem, the price of the... 相似文献
Currently, a huge amount of cargo is transported via containers by liner shipping companies. Under stochastic demand, repacking operations and carbon reduction, which may lead to an increase in effectiveness and environmental improvement, have been rarely considered in previous literature. In this paper, we investigate a container transshipment route scheduling problem with repacking operations under stochastic demand and environmental protection. The problem is a combinatorial optimization problem. Lacking historical data, a chance-constrained programming model is proposed to minimize the total operating and environment-related costs. We choose two distribution-free approaches, i.e., approximation based in Markov’s Inequality and Mixed Integer Second-Order Conic Program to approximate the chance constraints. As the loses induced by unfulfilled demand are not taken into account in the above model, a scenario-based model is developed considering the loses. Risk-neutral model may provide solutions that perform poorly while considering uncertainty. To incorporate decision makers’ perspectives, therefore, we also propose a risk-averse model adopting a risk aversion measure called Conditional Value-at-Risk to meet different preferences. Finally, we conduct computational experiments based on real data to compare the performances of the modeling methods and illustrate the impacts by testing different risk levels and confidence levels.