We say a divisien (S = SA or SB, g) is equal if it divides set S into two subsets SA and SB satisfying f (SA) = f (SB), where f is som attribute function on these sets and SA or SB need not be empty. In the present pa...We say a divisien (S = SA or SB, g) is equal if it divides set S into two subsets SA and SB satisfying f (SA) = f (SB), where f is som attribute function on these sets and SA or SB need not be empty. In the present paper, we have obtained some equal divisions on set of ordered tree with n modes: two equal divisions on tree set; two equal divisions on node set; two divisions on leaf set, one is equal, another one is ''asymptotically'' equal. We conclude that, ordered tree set is on its good behavior in equal dividing.展开更多
With the development of Internet, frequent pattern mining has been extendedto more complex patterns like tree mining and graph mining. Such applications arise in complexdomains like bioinformatics, web mining, etc. In...With the development of Internet, frequent pattern mining has been extendedto more complex patterns like tree mining and graph mining. Such applications arise in complexdomains like bioinformatics, web mining, etc. In this paper, we present a novel algorithm, namedChopper, to discover frequent subtrees from ordered labeled trees. An extensive performance studyshows that the newly developed algorithm outperforms TreeMiner V, one of the fastest methodsproposed previously, in mining large databases. At the end of this paper, the potential improvementof Chopper is mentioned.展开更多
In this paper, we consider testing the hypothesis that all multinomial populations in the stratified contingency table are identically distributed against the alternative that all these popula- tions are in simple tre...In this paper, we consider testing the hypothesis that all multinomial populations in the stratified contingency table are identically distributed against the alternative that all these popula- tions are in simple tree order. We provide an asymptotic represen- tation of the order-restricted maximum likelihood estimate of the unknown parameters. The resulting estimators are proven to be ~n-consistent and asymptotically normal under appropriate conditions. A chi-squared test method is used for this hypothesis test problem. A real data set is applied to illustrate our theoretical result.展开更多
基金Supported by China National Natural Science Foundation.
文摘We say a divisien (S = SA or SB, g) is equal if it divides set S into two subsets SA and SB satisfying f (SA) = f (SB), where f is som attribute function on these sets and SA or SB need not be empty. In the present paper, we have obtained some equal divisions on set of ordered tree with n modes: two equal divisions on tree set; two equal divisions on node set; two divisions on leaf set, one is equal, another one is ''asymptotically'' equal. We conclude that, ordered tree set is on its good behavior in equal dividing.
文摘With the development of Internet, frequent pattern mining has been extendedto more complex patterns like tree mining and graph mining. Such applications arise in complexdomains like bioinformatics, web mining, etc. In this paper, we present a novel algorithm, namedChopper, to discover frequent subtrees from ordered labeled trees. An extensive performance studyshows that the newly developed algorithm outperforms TreeMiner V, one of the fastest methodsproposed previously, in mining large databases. At the end of this paper, the potential improvementof Chopper is mentioned.
基金Supported by the National Natural Science Foundation of China (10771163)
文摘In this paper, we consider testing the hypothesis that all multinomial populations in the stratified contingency table are identically distributed against the alternative that all these popula- tions are in simple tree order. We provide an asymptotic represen- tation of the order-restricted maximum likelihood estimate of the unknown parameters. The resulting estimators are proven to be ~n-consistent and asymptotically normal under appropriate conditions. A chi-squared test method is used for this hypothesis test problem. A real data set is applied to illustrate our theoretical result.
