The NIST Statistics Reference Database (StRD) benchmarks are a collection of data sets, computational problems, and certified results that are designed to test the accuracy of statistical software. For general information on the datasets and computational problems, see:
We have tested RATS using both the linear regression and non-linear regression data sets, and present the results below.
These test the linear regression capabilities of statistical packages with problems that vary in the degree of computational difficulty.
To see how RATS performed on these tests, click on NIST Linear Regression Results.
For additional details on the Linear Regression benchmark data sets and programs, see:
The Nonlinear Regression benchmarks test the quality of a statistical package's non-linear estimation algorithms and numerical precision. For each program, the estimation is run using three different sets of initial conditions. The first set are intentionally set up to be "far" from the actual values (making the estimation process more difficult), the second set are closer to the actual values (which should make estimation easier), while the third set starts the estimation from the final estimated values (the challenge being for the algorithm to determine that it is already at a maximum).
To see how RATS performed on these tests, click on NIST Non-Linear Regression Results.
For more details on these problems, see: