The data in its raw form may not be of much use for the end customer. In the attempt to extract the knowledge from the data, the concept of data mining is extremely useful. This chapter explains how the data is to be filtered out to extract useful information. Often, exactly this information is requested by the players of the supply chain towards decision making. They are not interested in the binary data.
TopBackground
Data mining (Han, J., Kamber, 2001) involves the collection and analysis of a large volume of data that is generally not feasible to carryout manually and the help of machine intelligence is sought. Eg., analysis of the hourly variations in the stock prices of a company based on the data gathered over 15 days. This data will be helpful for predictions of the future values and decision making.
Data mining architecture comprises of (Mierswa, Ingo and Wurst, Michael and Klinkenberg, Ralf and Scholz, Martin and Euler, Timm, 2006) archival, retrieval, analysis and usage of the data. The algorithms and concepts borrowed from the Artificial intelligence, neural networks, Fuzzy sets etc are used for the realization of the same.
Data mining (Ethem Alpaydın, 2004) runs the statistics and machine learning algorithms over different data formats. While the data is getting translated in to usable knowledge, patterns will be identified in the data. The patterns enrich the knowledge model.
Knowledge fusion is an important part of knowledge management that consolidates and brings in the knowledge distributed across in to one integrated platform. Bayesian networks are proposed for the fusion of the knowledge. The usage of Neural Networks for Bayesian decision to facilitate data mining is the topic of discussion in this chapter. It brings in the required automation for information sorting, classification and clustering.