General Strategies for Data Cube computation in Data Mining

Pre-requisites: Data mining

Data Mining can be referred to as knowledge mining from data, knowledge extraction, data/pattern analysis, data archaeology, and data dredging. In data mining, a data cube is a multi-dimensional array of data that is used for online analytical processing (OLAP). 

Here are a few strategies for data cube computation in data mining:

1. Materialized view

This approach involves pre-computing and storing the data cube in a database. This can be done using a materialized view, which is a pre-computed table that is based on a SELECT statement. 

• Advantage: The advantage of this approach is that data cube queries can be answered quickly since the data is already pre-computed and stored in the database. 
• Disadvantage: The disadvantage is that the materialized view needs to be updated regularly to reflect changes in the underlying data.

2. Lazy evaluation

This approach involves delaying the computation of the data cube until it is actually needed.

• Advantage: The advantage of this approach is that it allows the data cube to be computed on-the-fly, which can be more efficient if the data cube is not needed very often.
• Disadvantage: The disadvantage is that data cube queries may be slower since the data cube needs to be computed each time it is accessed.

3. Incremental update

This approach involves computing the data cube incrementally, by only updating the parts of the data cube that have changed. 

• Advantage: The advantage of this approach is that it allows the data cube to be updated more efficiently since only a small portion of the data cube needs to be recomputed. 
• Disadvantage: The disadvantage is that it can be more complex to implement since it requires tracking changes to the data and updating the data cube accordingly.

4. Data cube approximation

This approach involves approximating the data cube using sampling or other techniques. 

• Advantage: The advantage of this approach is that it can be much faster than computing the data cube exactly.
• Disadvantage: The disadvantage is that the approximated data cube may not be as accurate as the exact data cube.
   

5. Data warehouse

A data warehouse is a central repository of data that is designed for efficient querying and analysis. Data cubes can be computed on top of a data warehouse, which allows for fast querying of the data. However, data warehouses can be expensive to set up and maintain, and may not be suitable for all organizations.

6. Distributed computing

In this approach, the data cube is computed using a distributed computing system, such as Hadoop or Spark. 

• Advantage: The advantage of this approach is that it allows for the data cube to be computed on a large dataset, which may not fit on a single machine.
• Disadvantage: The disadvantage is that distributed computing systems can be complex to set up and maintain, and may require specialized skills and resources.

7. In-memory computing

This approach involves storing the data in memory and computing the data cube directly from memory.

• Advantage: The advantage of this approach is that it allows for very fast querying of the data since the data is already in memory and does not need to be retrieved from disk.
• Disadvantage: The disadvantage is that it may not be practical for very large datasets, since the data may not fit in memory.

8. Streaming data

This approach involves computing the data cube on a stream of data, rather than a batch of data. 

• Advantage: The advantage of this approach is that it allows the data cube to be updated in real-time, as new data becomes available.
• Disadvantage: The disadvantage is that it can be more complex to implement, and may require specialized tools and techniques.

Note: Sorting, hashing, and grouping are techniques that can be used to optimize data cube computation, but they are not necessarily strategies for data cube computation in and of themselves.