The Motivations of Data Mining

The interest in data mining of researchers and practitioners with different backgrounds has increased steadily year after year. This growth is due to several reasons.

First, data mining plays today a fundamental role in analyzing and understanding the vast amount of information collected by business, government, and scientific applications. The ability to analyze large bodies of data and extract from them relevant knowledge has become a valuable service for most organizations that operate in the highly globalized and competitive business arena. The technical skills required to operate and put to use data-mining techniques are now appreciated, and often required, by the business intelligence units of financial institutions, government agencies, telecommunication companies, service providers, retailers, and distribution operators. A second reason is to be found in the excellent and constantly improving quality of the methods and tools that are being developed in this field. Advanced mathematical models, state-of-the-art algorithmic techniques, and efficient data management systems, combined with a decreasing cost of computational power and computer memory, are now able to support data analysts with methodologies and tools that were not available a few years ago. Furthermore, such instruments are often available at low cost and with easy-to-use interfaces, integrated into well-established data management systems.

A third reason that is not to be overlooked is connected with the role that data-mining methods are playing in providing support to basic research in many scientific areas. To mention an example, biology and genetics are currently enjoying the results of the application of advanced mining techniques that allow discovery of valuable facts in complex data gathered from experiments in vitro.

Finally, we wish to mention the impulse to methodological research that has been given in many areas by the open problems posed by data-mining applications. The learning and classification problems coming from real-life problems have been exploited through many mathematical theories under different formalizations, and theoretical results of unusual relevance have been reached in optimization theory, computer science, and statistics, also thanks to the many new and stimulating problems.