Design and Implementation of an Improved Data Warehouse on Clinical Data

Abstract

Data Warehouse is a repository to store huge detailed and summaries data for historical data analysis. In a decision support system which stores data from remote, complex and heterogeneous operational data sources . A clinical data warehouse contains complex, heterogeneous data from different data sources. In literature, there are different data warehouse architectures are present with there own design issues, which are relevant to different application areas. In this paper, we proposed a conceptual and logical view of data warehouse architecture along with physical implementation of the data warehouse. Our main focus in this paper is to efficiently handle the complex heterogeneous medical data stored into the warehouse and improve the performance of data warehouse for data analysis. Here, we proposed a partitioning concept of the dimension tables and fact tables for optimizing the response time, minimizing the disk IO, along with reducing the joining cost of the data warehouse. To show the effectiveness of our system, we, compare with different joining techniques of the dimension and fact tables of fact-consolidated data warehouse schema. A mathematical cost model of disk IO optimization is being calculated. SQL window partitioning techniques are being used for data analysis of the proposed data warehouse. After storing complex heterogeneous data in well organized and efficient way in a data warehouse, efficient searching techniques need to be incorporated. Here, bitmap indexing technique is used for the purpose.