What Converge is
In statistics, convergence is the process by which a sequence of estimates or values of a parameter approaches the true value of that parameter as the sample size gets larger. It is a fundamental concept in the study of sampling distributions and is an important consideration when choosing an appropriate sample size.
Converging can be a multi-step process and can include the following steps:
- Estimate a parameter with a sample of a given size.
- Increase the sample size and re-estimate the parameter.
- Compare the two estimates.
- If the estimates are not close enough, repeat the steps until the difference between the estimates is small enough for the desired level of accuracy.
- At this point, the sequence of estimates has converged, and the estimates are considered to be close enough for the desired level of accuracy.
Examples
-
Converge is used in statistics when performing a simulation to ascertain the accuracy of a statistical model.
-
Converge is used to determine the point of stability in a mathematical function or equation by plotting points along the curve and observing their behavior as the number of iterations increases.
-
Converge is used in regression to indicate how closely the model fits the data, by comparing the mean squared error of the fitted line to the mean squared error of the data itself.