Horizontal Aggregation Approaches for Data Analysis

ABSTRACT Data mining is used to analyze the data which is derive from different perspectives and then summarize into useful information, this data is then used by the different users for analyzing the data as well as for preparation of data sets. A data set is collection of data and generally presented in the tabular form. Usually preparation of data set is most time consuming as it involves complex SQL queries, joining tables and aggregate functions. Traditional RDBMS represent tables with vertical format and returns one number per row which means that returns a single value output which is not suitable for preparing a data set. This paper mainly focuses on various horizontal aggregation methods and k means clustering algorithm which is used to partition data sets after horizontal aggregations. Horizontal Aggregation consists of three methods that are SPJ, CASE and PIVOT methods. Result of horizontal aggregations which are obtained from these methods results in large set of data, partition it into homogeneous clusters is important part in the system. k means clustering algorithm is best suited to implement that.