Levels of data mining operations
The aggregate or the Macro level
Mining at the macro level gives us a broad overview of the data e.g. when customer of the retail store are segmented by profitability criteria, we obtain clusters who are profitable to various extent. Knowledge obtained by mining at macro level is useful when dealing with situations where:
• We are dealing with a customer about whom we do not have individual information. Hence, we need to extrapolate the characteristics of the group to which he/she might belong. In retail store example, a store can segment its customers on basis of age and characteristics can be extracted. When a new customer enters the store, the salesman could use his intuition in arriving at the customer’s age and recover the characteristics of that age group such as the frequently bought products, colour preferences, etc.
• Targeting new set of customers. If the retail chain has opened a new store it can use the data from the most similar current store to predict the behavior of the new prospects.
• We are dealing with aspects of the service, which influence a majority of the customer and therefore cannot be customized to suit individual tastes, example being the design of the physical layout of a retail store.
• Predicting the possibility of an action that the cu has never undertaken. A customer might not have tried out a new product because he/she was not aware of it. A salesman can encourage him/her to try out the product if his/her profile matches that of the current product users.
The aggregate or the Macro level
Mining at the macro level gives us a broad overview of the data e.g. when customer of the retail store are segmented by profitability criteria, we obtain clusters who are profitable to various extent. Knowledge obtained by mining at macro level is useful when dealing with situations where:
• We are dealing with a customer about whom we do not have individual information. Hence, we need to extrapolate the characteristics of the group to which he/she might belong. In retail store example, a store can segment its customers on basis of age and characteristics can be extracted. When a new customer enters the store, the salesman could use his intuition in arriving at the customer’s age and recover the characteristics of that age group such as the frequently bought products, colour preferences, etc.
• Targeting new set of customers. If the retail chain has opened a new store it can use the data from the most similar current store to predict the behavior of the new prospects.
• We are dealing with aspects of the service, which influence a majority of the customer and therefore cannot be customized to suit individual tastes, example being the design of the physical layout of a retail store.
• Predicting the possibility of an action that the cu has never undertaken. A customer might not have tried out a new product because he/she was not aware of it. A salesman can encourage him/her to try out the product if his/her profile matches that of the current product users.