New divergence measures improve accuracy of goodness of fit tests.
The article explores new ways to test how well data fits a model using different measures of divergence. By introducing a family of divergence measures with a parameter "a", the researchers found optimal values of "a" to make these tests more accurate.