Abstract: Equipment support capability is an important constituent part of army combat power. For the sake of improving the accuracy of predicting the equipment support capability, an improved parallel prediction model is proposed for the problems existing in current prediction algorithms and their combined models. The indexes and weights of equipment support capability are confirmed by text mining. A non-isometric multi-scale interval fuzzy time series model is established by modifying the interval scale algorithm. At the same time, the particle swarm optimization algorithms about particle speed, location and inertia weight value are improved to optimize the parameters of support vector machine. An improved parallel prediction model is constructed based on the above two models, by which the weight values are calculated and the predicted weight values and the predicted values are combined to obtain the final predictive value. The given example shows that the improved prediction model is accurate.

Key words: ordnance science and technology, equipment support capability, parallel prediction, time series, support vector machine, particle swarm optimization algorithm