Weighted linear regression for symbolic interval-values data with outliers

In many real word applications, engineers are often asked to deal with symbolic interval-values data, which may also contain outliers. To overcome the problem caused by those outliers in the symbolic interval-value data, the weighted linear regression methods are applied. These weighted linear regressions are to extend the Centre method, the MinMax method and the Centre and Range method, respectively to the robust version of the linear regression methods with outliers. From the simulation results, the proposed methods are superior to their non-robust versions of the linear regression methods.

Перевод пока недоступен