Optimal data warehouse design with data marts and data cube aggregation

Almost all Data Mining techniques and algorithms suffer from the high time complexity due to the huge amount of data and the algorithms nature. In general it can be concluded that the time complexity of different mining algorithms is a function of number of records in the dataset, number of features and number of distinct values in these features in addition to some other factors. At the same time data privacy, ownership and security represent a big challenge to the warehouse projects. The work in this paper tends to improve such complexity and reduce the required accessing time for different queries and improve privacy, security, data ownership throughout a combination of warehouse design with many fact tables (Galaxy model) and using data cube approach containing both detailed and highly summarized data. Cause of death free dataset available on the internet that include more than 14 million records was used to study the effectiveness of the proposed approach. The results obtained showed high improvement of the mentioned criteria for the proposed approach compared with the traditional way of dealing with such data.