Pattern detection and discovery

Viac o knihe

The growth of large electronic databases has spurred interest in methods for uncovering structures within them, commonly referred to as data mining. A key aspect of this field focuses on identifying and detecting anomalous, interesting, or valuable records, known as patterns. Examples include detecting fraud in credit-card transactions, identifying coincident purchases in supermarkets, analyzing significant nucleotide sequences in genetics, and recognizing characteristic traces in EEG records. While tools for pattern detection have been developed within the data mining community, they have also emerged in various other research fields, often without recognizing the shared underlying problems. Each discipline has its own extensive and rapidly growing literature, making it challenging to stay updated across multiple areas, especially when they use unfamiliar technical language. Consequently, valuable opportunities may be missed as discoveries made in one field do not transfer to others. This can lead to the rediscovery of breakthroughs or prolonged periods before solutions are found, resulting in wasted effort and lost opportunities.