论文部分内容阅读
通过提取陶瓷样品对瓷器鉴定的影响因素,采用了基于欧氏距离分析法,得到了各个因素的权重,从而能够确定影响因素的关键因素.通过关键因素作为神经网络模型输入变量,建立基于提取权重的概率神经网络算法.实例分析表明:算法通过提取权重能够提高分类准确度.在大样本实例分析中,算法比其他统计分析和相似度算法具有更快的收敛速度,并能够应用到其它数据处理中,具有广泛的适用性.
Based on the Euclidean distance analysis method, the weight of each factor was obtained by extracting the influence factors of ceramic samples on the porcelain identification, and the key factors of influencing factors were determined.According to the key factors as input variables of neural network model, .Example analysis shows that the algorithm can improve the classification accuracy by extracting weights.Compared with other statistical analysis and similarity algorithms, the algorithm can converge faster than other statistical analysis and similarity algorithms in large sample case analysis and can be applied to other data processing In a wide range of applicability.