Extracting knowledge from data
Knowledge extraction from data (KDE)

The modern approach to extracting knowledge from data is intended to be as general as possible. It favors neither a particular source of information (which may be locally stored or distributed), nor a specific type of data (which may be structured as attribute-values, texts of varying lengths, images or video sequences). It is not limited to the latest analysis tools, and explicitly incorporates methods for data preparation, analysis and validation of the knowledge produced. Most of these methods come from the fields of statistics, data analysis, machine learning and pattern recognition.

EDC is an anthropocentric process: the knowledge extracted must be as intelligible as possible for the user. It must be validated, formatted and organized. Let's take a closer look at all these notions and situate them within the overall EDC process.

EDC...