Business managers accounting for commercial success or non-success of the organization have to gain knowledge needful for correct decision acceptance. These knowledge represent sophisticated information hidden in enterprise data. One possibility, how to extract mentioned knowledge from data, is to use so-called datamining assets.The paper deals with an application of chosen basic methods of knowledge discovering in databases for area of customer-provider relation and it presents, how to avail acquired knowledge as basis of managerial decisions leading to improving of customer relationship management. It solves prediction, whose aim is, on the basis of some attributes of exploring objects, to predict future behaviour of objects with these attributes. This way acquired knowledge, as the output of prediction, then can markedly help competent enterprise manager with planning of marketing strategies, for example so-called cross-selling and up-selling. The contribution describes a whole operation of available data processing: from its purifying, over its preparation for mining task, to self processing by the help of SAS Enterprise Miner tool. Regression analysis, neural network and decision tree, whose principles are briefly explained in this paper too, were used for knowledge mining. The estimation of customer behaviour was tested by two mining task varying in attribute using and in categories number of one of predicive attributes. The results of these two tasks are confronted by the help of prediction fruitfulness charts.
Customer Relationship Management: Related Theories, Challenges and Application in Banking Sector
