中文摘要：

随着汽车的普及应用，座椅舒适性成为关注焦点。然而，目前有关汽车座椅舒适度评价主观性较大，缺乏客观可行的评价方法。笔者通过有针对性的客户问卷调查，应用人工神经网络技术，以座椅的各部分特征性能为输入变量，建立了汽车座椅舒适度的人工智能评价系统。此外，为了进一步提高该系统的预测精度，笔者还应用了聚类分析方法。通过样本类数的调整，能将系统预测误差从原来的41．09％最大降至27．845％。

英文摘要：

With the wide application of automobiles, automobile seat comfort has attracted the attention of researchers. However, there is not an objective and feasible evaluation method. In this paper, based on a survey of automo- bile customers, an effective automobile seat comfort evaluation system is constructed by using artificial neural network; its input variables are characteristics of different seat parts. In addition, in order to improve the prediction accuracy of the system, cluster analysis method is applied. By modif）;ing sample cluster number, the prediction error can be reduced from 41.09% to 27. 845%.