Published by World Academic Press, World Academic Union ISSN 1746-7659, England, UK Journal of Information and Computing Science Vol. 6, No. 4, 2011, pp. 269-278 A Framework for Reducing Multidimensional Database to Two Dimensions Adio Akinwale 1 , Kolawole Adesina 2 and Olusegun Folorunso 3 Department of Computer Science, University of Agriculture, Abeokuta, Nigeria (Received April 8, 2011, accepted May 17, 2011) Abstract. This work used a method of Matrix Decomposition Algorithm to obtain a new dataset of genetic epistasis as a surrogate for a multidimensional dataset which transformed multidimensional database to a 2- dimensional database. It employed decomposition algorithms based on Boyce Codd Normal Form for minimizing anomalies. The decomposition and reversible algorithms were used on relationship among object attributes and were implemented. The implemented program ran on sample genetic epistasis datasets of up to 10 dimensions and it was shown that multidimensional datasets can be reduced to two dimensions. It was established that the time taken to generate a sequence of tuples from multidimensional database to a 2- dimensional dataset was directly proportional to the number of genes considered. The result showed that the reduced 2-dimensional database did not require any in-built functions which take long processing time for generating query result as against querying of multidimensional dataset. The reduced 2-dimensional dataset was reversible to the original multidimensional dataset for lossless join operation which indicated that there was no loss of data values or tuple. The method was compared with existing reduction techniques and it was found that data access was very fast with decomposition algorithm than relational model. Keywords: Matrix decomposition algorithm, multidimensional database, genetic epistatis, principal component analysis, project pursuit method, relational model 1. Introduction Nowadays, emerging data are multidimensional in nature. Multidimensional database technology is being applied to distributed data and to new types of data that current technology often cannot adequately analyze. The global reason for the multidimensional database’s rise is to facilitate flexible, high performance access and analysis of large volumes of complex and interrelated data [24]. Multidimensional database structure has evolved to match closely the way people visualize data [17]. Thus, people think of their businesses in multidimensional terms. For example, a sales manager will want to know the sales of a particular product over a period of time and may be in a particular region or location. This is a multidimensional view of sales data. The higher the dimensionality, the higher the volume of the data, and consequently, the more complex the data is to manage. To lessen the scalability problem of processing large datasets, data reduction technique is proffered. Data reduction techniques are useful to reduce the scalability problem of processing large datasets. In order to efficiently archive and process growing multidimensional datasets, the work presents and compares reduction and reversibility technique algorithms that reduce multidimensional database to a two dimensional database and also reverse the two dimensional back to its original form without loss of data. 2. Literature Review Multidimensional database is stored in such a way as to be represented to the user as a hypercube or multidimensional array, where each core value or fact occupies a cell indexed by a unique set of dimension values. Agrawal et. al. also asserted that multidimensional database is a key technology in the enabling of interactive analyses of large amounts of data for decision-making purposes [1]. It is used to process and analyze large and complex datasets. Multidimensional database originated from the multidimensional matrix 1 E-mail address: aatakinwale@yahoo.com; 2 E-mail address: kolawole_adesina@yahoo.co.uk; 3 E-mail address: folorunsolusegun@yahoo.com