S.Muthukumaran, A. Victoria Anand Mary, M.Valli
Data preprocessing is a vital step in data mining. Preprocessing resolves various types of data discrepancies encountered in large databases in order to produce quality data for the mining task. Data preprocessing includes four fundamental steps namely data cleaning, integration, reduction and transformation. There are various techniques involved in each step of Data preprocessing. In order to develop quality data, a data miner must decide the most appropriate techniques in every step of preprocessing. In this paper we focus on data reduction, particularly data discretization as one of the most important preprocessing step. Data reduction involves reducing the data distribution by reducing the range of continuous data into a range of values or categories. Data discretization plays a major role in reducing the attribute intervals of data values. Finding an appropriate number of discrete values will improve the performance of data mining modeling, particularly in terms of classification accuracy. This research paper proposes four levels of data discretization taxonomy as follows namely (i) hierarchical and non-hierarchical; (ii) splitting, merging, and combination; (iii) supervised and unsupervised combinations; (iv) binning, entropy and chi-square merge techniques
Data Preprocessing, Data reduction, Data Discretization.