Community Research and Development Information Service - CORDIS

Improved knowledge from data mining

Today's enterprises accumulate vast databases that contain a wealth of important data, however, due to their considerable volume this information remains largely inaccessible. In order to utilise such a valuable source of information, new data mining techniques are being developed to convert data into useful information.
Improved knowledge from data mining
Data mining is a technique, which allows large databases to be searched in order to find hidden patterns and trends, which would otherwise remain buried in the vast amounts of data stored. These trends can hold vital information on a multitude of subjects such as manufacturing-related knowledge or consumer-buying habits.

There are three stages in the process, namely, exploration, pattern identification with validation, and deployment. Exploration, the first stage, involves preparing the data stored in the database. This may involve data cleaning, data transformation and dividing data into sets to make the information more manageable. The second stage involves building a model, which can predict the results contained in the database and explain variations in results. In the verification part of stage two, this model has to be tested over a number of samples to ensure that it gives consistent results. Deployment, the final stage, applies the model to new data in order to generate predictions.

The technique is particularly useful in producing long term strategic plans for manufacturing industries from existing performance knowledge. This data mining terminology is now being developed further. While, methods of developing improved data mining models are being explored. Other areas being addressed include mining databases, which contain past operational history and performance for manufacturing enterprises. Mining databases use data mining to determine requirements, aims and objectives for new manufacturing system designs and developing new applications in manufacturing system design.

New methods in data mining will enable manufacturing industries to exploit their own databases, to obtain a valuable record of their business history and allow long-term objectives to be met.
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top