不同数据预处理对雷达点迹分类的影响

Influence of Different Data Preprocessing on Radar Plot Classification

利用全连接神经网络研究雷达点迹分类问题时,在模型训练前需要对打上标签的点迹数据进行预处理,不同的数据预处理方法,会影响模型训练的效果。提出使用一种最大最小值联合处理的方法,与常用的其它4种不同数据预处理方法进行对比,分析不同数据预处理方法对雷达点迹分类效果的影响。本文所用方法明显优于常用方法,不仅加快了网络的学习训练速度,也提高了模型的分类精度。

When using the fully connected neural network to study the radar plot classification problem,it is necessary to preprocess the labeled plots before model training.Different data preprocessing methods will affect the effect of model training.This paper proposes a joint processing method of maximum and minimum value,which is compared with other four different data preprocessing methods,and analyzes the influence of different data preprocessing methods on the effect of radar plot classification.The method used in this paper is obviously superior to the common methods,which not only speeds up the learning and training speed of the network,but also improves the classification accuracy of the model.