Abstract: Analyzes the basically methods that radar chart presents multi-dimensional data，and then introduces feature selection and information fusion technique based on radar chart feature.A novel method based on chart feature for designing the classifier of multi-dimensional data is proposed，which uses radar chart to map.This method is also simple and intuitionistic.Multi-dimensional data and applied fuzzy rules to recognize the radar chart in order to achieve automatic classification.Different category multi-dimensional data forms different radar chart.Radar chart feature shows the mostly data feature.The result of experiment presents better result of category and precision.

Key words: multi-dimensional data, radar chart, classifier, feature selection, information fusion

摘要： 分析介绍了用雷达图表示多维数据以及雷达图的图形特征选取和融合的基本方法，提出了一种基于多维数据雷达图表达的可视化图形分类新方法，该方法用雷达图表示多维数据，不同类别的多维数据对应不同的雷达图形，形成以雷达图形特征为表达主要特征的分类方法。并通过模糊推理方来来自动识别雷达图形，完成自动分类。实验表明，此方法简单、直观，实现了分类过程可视化、分类结果可视化，并且有良好的分类效果。

关键词: 多维数据, 雷达图, 分类器, 特征选择, 信息融合