Abstract: To deal with combat effectiveness prediction of the military system, a SVR-based crucial evaluation indexes mining method was carefully investigated. The key indexes in the effectiveness evaluation were found by comparing partial derivatives. The model of efficiency prediction based on Elman neural networks, which used effective optimized indexs and values as input and output, was exerted to combat effectiveness prediction of C⁴ISR. The results show that the method can reduce the complexity of prediction model, and avoid uncertain factors existing in system, which provide effective technical support for the combat effectiveness prediction scientifically.

Key words: combat effectiveness, Elman neural networks, SVM, effectiveness prediction model