The rollup and the drilldown are quite frequent based on dimension hierarchies in on-line analytical processing （OLAP）, but prefixCube does not directly support dimension hierarchies. The PrefixCube is extended for incorporating hierarchical data cubes, i.e. cubes with hierarchical dimensions ,thus obtaining HierPrefixCube. HierPrefixCube retains advantages on computation and organization of PrefixCube, and it can directly support aggregate queries on levels of the dimension hierarchy.
This paper combines computational intelligence tools: neural network, fuzzylogic, and genetic algorithm to develop a data mining architecture (NFGDM), which discovers patternsand represents them in understandable forms. In the NFGDM, input data arepreprocesscd byfuzzification, the preprocessed data of input variables arc then used to train a radial basisprobabilistic neural network to classify the dataset according to the classes considered, A ruleextraction technique is then applied in order to extract explicit knowledge from the trained neuralnetworks and represent it m the form of fuzzy if-then rules. In the final stage, genetic algorithmis used as a rule-pruning module to eliminate those weak rules that are still in the rule bases.Comparison with some known neural network classifier, the architecture has fast learning speed, andit is characterized by the incorporation of the possibility information into the consequents ofclassification rules in human understandable forms. The experiments show that the NFGDM is moreefficient and more robust than traditional decision tree method.
There exists an inherent difficulty in the original algorithm for the construction of Dwarf, which prevents it from constructing true Dwarfs. We explained when and why it introduces suffix redundancies into the Dwarf structure. To solve this problem, we proposed a completely new algorithm called PID. It bottom-up computes partitions of a fact table, and inserts them into the Dwarf structure. If a partition is an MSV partition, coalesce its sub-Dwarf; otherwise create necessary nodes and cells. Our performance study showed that PID is efficient. For further condensing of Dwarf, we proposed Condensed Dwarf, a more com- pressed structure, combining the strength of Dwarf and Condensed Cube. By eliminating unnecessary stores of “ALL” cells from the Dwarf structure, Condensed Dwarf could effectively reduce the size of Dwarf, especially for Dwarfs of the real world, which was illustrated by our experiments. Its query processing is still simple and, only two minor modifications to PID are required for the construction of Condensed Dwarf.
