An approach to benefit from data mining in business environments. (dissertation)

The amount of stored data increases drastically and at high speed. Scanner-cashier-systems, production-control systems and others generate millions of data items each day. Since information has become an essential resource for companies, the analysis of such data gets more and more important. Traditional methods are of little help when such a huge amount of data should be analyzed. The over-abundance of irrelevant data, which was already a topic in the 60ies, becomes more and more a severe problem. Therefore, the need for methods is increasing, which automatically filter interesting patterns out of large databases. With the goal to develop such methods the research area of data mining has established itself in the meantime. Research activities are reported in many areas and the improvements achieved seem promising. In business environments available systems are currently introduced to the market, but still have to prove their usability. Such systems usually don´t rely on an own method or algorithm, but by combining well-known methods information filters are constructed. But still, the current systems have only little in common with the ideal system, which automatically identifies useful patterns out of any database. Therefore, it is absolutely important for a user to know, how powerful those systems are and what strengths and weaknesses they include