Combining the advantages of the stratified sampling and the importance sampling, a stratified importance sampling method (SISM) is presented to analyze the reliability sensitivity for structure with multiple failure...Combining the advantages of the stratified sampling and the importance sampling, a stratified importance sampling method (SISM) is presented to analyze the reliability sensitivity for structure with multiple failure modes. In the presented method, the variable space is divided into several disjoint subspace by n-dimensional coordinate planes at the mean point of the random vec- tor, and the importance sampling functions in the subspaces are constructed by keeping the sampling center at the mean point and augmenting the standard deviation by a factor of 2. The sample size generated from the importance sampling function in each subspace is determined by the contribution of the subspace to the reliability sensitivity, which can be estimated by iterative simulation in the sampling process. The formulae of the reliability sensitivity estimation, the variance and the coefficient of variation are derived for the presented SISM. Comparing with the Monte Carlo method, the stratified sampling method and the importance sampling method, the presented SISM has wider applicability and higher calculation efficiency, which is demonstrated by numerical examples. Finally, the reliability sensitivity analysis of flap structure is illustrated that the SISM can be applied to engineering structure.展开更多
In recent years, the deep web has become ex- tremely popular. Like any other data source, data mining on the deep web can produce important insights or summaries of results. However, data mining on the deep web is cha...In recent years, the deep web has become ex- tremely popular. Like any other data source, data mining on the deep web can produce important insights or summaries of results. However, data mining on the deep web is chal- lenging because the databases cannot be accessed directly, and therefore, data mining must be performed by sampling the datasets. The samples, in turn, can only be obtained by querying deep web databases with specific inputs. In this pa- per, we target two related data mining problems, association mining and differential rule mining. These are proposed to ex- tract high-level summaries of the differences in data provided by different deep web data sources in the same domain. We develop stratified sampling methods to perform these min- ing tasks on a deep web source. Our contributions include a novel greedy stratification approach, which recursively pro- cesses the query space of a deep web data source, and con- siders both the estimation error and the sampling costs. We have also developed an optimized sample allocation method that integrates estimation error and sampling costs. Our ex- perimental results show that our algorithms effectively and consistently reduce sampling costs, compared with a strat- ified sampling method that only considers estimation error. In addition, compared with simple random sampling, our al- gorithm has higher sampling accuracy and lower sampling costs.展开更多
基金National Natural Science Foundation of China (10572117,10802063,50875213)Aeronautical Science Foundation of China (2007ZA53012)+1 种基金New Century Program For Excellent Talents of Ministry of Education of China (NCET-05-0868)National High-tech Research and Development Program (2007AA04Z401)
文摘Combining the advantages of the stratified sampling and the importance sampling, a stratified importance sampling method (SISM) is presented to analyze the reliability sensitivity for structure with multiple failure modes. In the presented method, the variable space is divided into several disjoint subspace by n-dimensional coordinate planes at the mean point of the random vec- tor, and the importance sampling functions in the subspaces are constructed by keeping the sampling center at the mean point and augmenting the standard deviation by a factor of 2. The sample size generated from the importance sampling function in each subspace is determined by the contribution of the subspace to the reliability sensitivity, which can be estimated by iterative simulation in the sampling process. The formulae of the reliability sensitivity estimation, the variance and the coefficient of variation are derived for the presented SISM. Comparing with the Monte Carlo method, the stratified sampling method and the importance sampling method, the presented SISM has wider applicability and higher calculation efficiency, which is demonstrated by numerical examples. Finally, the reliability sensitivity analysis of flap structure is illustrated that the SISM can be applied to engineering structure.
文摘In recent years, the deep web has become ex- tremely popular. Like any other data source, data mining on the deep web can produce important insights or summaries of results. However, data mining on the deep web is chal- lenging because the databases cannot be accessed directly, and therefore, data mining must be performed by sampling the datasets. The samples, in turn, can only be obtained by querying deep web databases with specific inputs. In this pa- per, we target two related data mining problems, association mining and differential rule mining. These are proposed to ex- tract high-level summaries of the differences in data provided by different deep web data sources in the same domain. We develop stratified sampling methods to perform these min- ing tasks on a deep web source. Our contributions include a novel greedy stratification approach, which recursively pro- cesses the query space of a deep web data source, and con- siders both the estimation error and the sampling costs. We have also developed an optimized sample allocation method that integrates estimation error and sampling costs. Our ex- perimental results show that our algorithms effectively and consistently reduce sampling costs, compared with a strat- ified sampling method that only considers estimation error. In addition, compared with simple random sampling, our al- gorithm has higher sampling accuracy and lower sampling costs.