针对K最近邻算法难以建立高精度的非线性模型问题,提出了一种基于特征加权的K最近邻预测方法。为提高模型的预测精度,该方法从特征重要程度的角度出发,采用Bootstrap特征加权方法对K最近邻算法进行特征加权。为了验证该方法的有效性,对无人机武器发射过程参数进行了预测。实验结果表明,与其它算法相比,该算法不仅体现了样本数据在模型中的作用,而且具有较高的预测精度。

A novel algorithm based on feature weighted KNN was proposed to solve the problem that it is hard to build nonlinear system modeling for KNN algorithm with high predicting precision. In order to improve prediction accuracy of the model, the Bootstrap feature weight method was used in KNN algorithm in view of feature importance of samples. The proposed method predicts UAV missile launch parameter extraction to verify effectiveness of the method. The simulation results show that the proposed algorithm not only shows role of samples in the model, but also has the advantages of high prediction accuracy compared with the other methods.