What are different approaches for software test estimation?

The best approach to software test estimation depends highly on a particular organization and the project and the experience of the personnel who are involved. Consider two projects of same size and complexity: one is life critical medical equipment and second was low cost computer game. In this, the appropriate test effort for medical equipment software is very large compared to the other one.

Some approaches that can be considered are:
- METRICS BASED APPROACH
This approach focuses on collecting the data for various projects of the organization and then this information can be used for any future test project planning. The expected required test time can be adjusted based on metrics or other information that is available.
- IMPLICIT RISK CONTEXT APPROACH
This approach focuses on using implicitly the risk context by a QA manager or project manager in combination with the past experiences to choose level of resources to allocate to testing. It is an intitutive guess based on experience.
- ITERATIVE APPROACH
This approach focuses on making an initial rough estimate. A refined estimate is made once the testing begins and after a small percentage of first estimate's work is done. The test plans can be refactored and a new estimate can be made. Repeat the cycle as and when necessary.
- TEST WORK BREAKDOWN APPROACH
This approach focuses on beaking the expected testing tasks into smaller tasks for which estimates can be made with reasonable accuracy. One point that has to be kept in mind is that an accurate and predictable breakdown for testing tasks is poosible.
- PERCENTAGE OF DEVELOPMENT APPROACH
This approach focuses on an estimation method for testing based on estimated programming effort.This method depends on project to project variations in risk, personnel, application types, complexity levels.