Data mining creates few model to identify patterns on attributes which is inscribed on dataset. Some of those patterns will create “descriptive” results, whereas some pattern creates “Predictive” results.
Generally, data mining identifies four kind of main patterns:
- Associations : groups same or similar events, for example a purchase of same item at supermarket.
- Predictions : forecast future events based on historical record. For example: predict future world cup winner based on winning streak pattern.
- Clusters : groups objects based on known characters, for example create a customer differentiation based on purchase history.
- Sequential relationships : find a relation between events, for example : predict a customer whom purchased a car will purchase spare parts a year later.
These general patterns are extracted from data manually for hundred years, but increased data volume on modern times demands an automated approach. The automated (or semi – automated) approach which analyzes very large amount of dataset is called Data Mining.
