Aggregation is the combination of two or more objects into a single object. Data aggregation is very useful when on the data set there are a number of values in the actual feature of a group, which will not deviate from the description on the feature if the value is merged. Aggregations that can be done are SUM, Average, MIN, and MAX. For example the Data purchase transaction in several branch distributors. Every day each branch does a lot of transactions. All transactions from each branch will provide large and complex transaction data. The transaction data will be simpler and do not deviate from the data descriptions if presented in a combined form each day in each branch. That way, data processing in data mining will be relatively simpler and faster computing.
As with customer purchase transaction data, it appears that every day there are a number of transactions for each branch. Views will be better and simpler if aggregated on total purchases for each day so that columns can be eliminated and memory usage for fewer tables. For example, we use the SUM aggregation on the column.
There are several reasons why aggregation should be done. The first reason is that smaller data sets require less storage memory, and processing time in the data mining algorithm becomes faster. As a result, these data sets can be processed with expensive algorithms in computing. The second reason is that aggregation acts to change the way views of data from low level to high level. The third reason is that the behavior of grouping objects or attributes is often more stable than the individual objects or attributes themselves.
As with customer purchase transaction data, it appears that every day there are a number of transactions for each branch. Views will be better and simpler if aggregated on total purchases for each day so that columns can be eliminated and memory usage for fewer tables. For example, we use the SUM aggregation on the column.
There are several reasons why aggregation should be done. The first reason is that smaller data sets require less storage memory, and processing time in the data mining algorithm becomes faster. As a result, these data sets can be processed with expensive algorithms in computing. The second reason is that aggregation acts to change the way views of data from low level to high level. The third reason is that the behavior of grouping objects or attributes is often more stable than the individual objects or attributes themselves.
Advertisement
No comments