摘要
针对部队平时弹药训练消耗量预测过程中，样本采集数目较少的实际情况，采用了一种新的预测方法——支持向量机。该方法基于统计学习理论的原理，较好地解决了小样本的学习问题。并以某部队1997-2002年弹药训练消耗量为学习样本，建立了弹药年消耗量的预测模型。计算结果表明，这种方法比传统的方法有更少的误差和更好的预测精度。
Based on the situation that the number of test samples is few in the course of predicting annual ammunition consumption level in training, a new method, support vector machine is given. The algorithm is based on statistical theory. It can better solve learning problem of small sample. By using the historical statistical data of ammunition consumption level from 1997 to 2002 as learning sample, a model is built to predict ammunition consumption level. The results show that the method can bring less error and better predicted precision compared with the traditional methods.
出处
《军械工程学院学报》
2006年第1期43-45,共3页
Journal of Ordnance Engineering College
基金
军队科研计划项目
关键词
支持向量机
弹药消耗量
预测
样本
support vector machine
ammunition consumption level
predict
sample
