基于改进的局部保持投影的战时备件分类

为提升战时合成部队备件保障效能，需对其进行有效分类，以便开展备件的预储预置。针对备件种类多、时效性强、影响分类因素复杂的现实问题，提出了基于改进的局部保持投影的备件分类方法。首先，根据战时备件分类储备的影响因素，作为备件分类的特征指标，其次，利用改进的局部保持投影的降维方法对备件原始特征数据进行特征降维，得到低维特征向量。再利用支持向量机（support vector machine，SVM）的分类器对低维数据进行分类。并通过量子粒子群对SVM的核函数参数进行寻优，提升备件分类精度，得到满足备件分类准确率最优时的降维维数和分类器参数。最后，通过对演习装备备件分类的实例分析，验证了模型的可行性和合理性，并对比分析了其他分类方法，表明该方法能够较好地解决战时备件分类的问题。

Classification of spare parts based on improved local preserving projection in wartime

In order to improve the support efficiency of combined army spare parts in wartime, it is necessary to classify them effectively in order to carry out the pre-storage and pre-configuration of spare parts. Aiming at the practical problems of many kinds of spare parts, strong timeliness and complex factors affecting the classification, a method of spare parts classification based on improved local preservation projection is proposed. Firstly, according to the influencing factors of the classified reserve of wartime spare parts, it is taken as the characteristic index of the spare parts classification. Secondly, the dimension reduction method of improved local preserving projection is used to reduce the dimension of the original feature data of spare parts, and low-dimensional feature vectors are obtained. Then the classifier of the support vector machine (SVM) is used to classify the low dimensional data. The kernel function parameters of the SVM are optimized by quantum particle swarm optimization to improve the accuracy of spare parts classification. The dimension reduction and classifier parameters are obtained when the spare parts classification accuracy is optimized. Finally, the feasibility and rationality of the model are verified by an example of the classification of spare parts of military manoeuvre equipment. By comparing and analyzing other classification methods, it shows that the method can better solve the problem of spare parts classification in wartime.